influx/telegraf
1
0
Fork 0
Code Issues Activity

Adding upstream version 1.34.4.

Browse source 
Signed-off-by: Daniel Baumann <daniel@debian.org>
This commit is contained in:
Daniel Baumann 2025-05-24 07:26:29 +02:00
parent e393c3af3f
commit 4978089aab
Signed by: daniel
GPG key ID: FBB4F0E80A80222F
4963 changed files with 677545 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

350
plugins/parsers/binary/README.md Normal file
View file

@ -0,0 +1,350 @@
# Binary Parser Plugin
The `binary` data format parser parses binary protocols into metrics using
user-specified configurations.
## Configuration
```toml
[[inputs.file]]
files = ["example.bin"]
## Data format to consume.
## Each data format has its own unique set of configuration options, read
## more about them here:
## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
data_format = "binary"
## Do not error-out if none of the filter expressions below matches.
# allow_no_match = false
## Specify the endianness of the data.
## Available values are "be" (big-endian), "le" (little-endian) and "host",
## where "host" means the same endianness as the machine running Telegraf.
# endianness = "host"
## Interpret input using the specified encoding
## Available values are "none" (raw bytes), "hex" and "base64"
# binary_encoding = "none"
## Multiple parsing sections are allowed
[[inputs.file.binary]]
## Optional: Metric (measurement) name to use if not extracted from the data.
# metric_name = "my_name"
## Definition of the message format and the extracted data.
## Please note that you need to define all elements of the data in the
## correct order with the correct length as the data is parsed in the order
## given.
## An entry can have the following properties:
## name -- Name of the element (e.g. field or tag). Can be omitted
## for special assignments (i.e. time & measurement) or if
## entry is omitted.
## type -- Data-type of the entry. Can be "int8/16/32/64", "uint8/16/32/64",
## "float32/64", "bool" and "string".
## In case of time, this can be any of "unix" (default), "unix_ms", "unix_us",
## "unix_ns" or a valid Golang time format.
## bits -- Length in bits for this entry. If omitted, the length derived from
## the "type" property will be used. For "time" 64-bit will be used
## as default.
## assignment -- Assignment of the gathered data. Can be "measurement", "time",
## "field" or "tag". If omitted "field" is assumed.
## omit -- Omit the given data. If true, the data is skipped and not added
## to the metric. Omitted entries only need a length definition
## via "bits" or "type".
## terminator -- Terminator for dynamic-length strings. Only used for "string" type.
## Valid values are "fixed" (fixed length string given by "bits"),
## "null" (null-terminated string) or a character sequence specified
## as HEX values (e.g. "0x0D0A"). Defaults to "fixed" for strings.
## timezone -- Timezone of "time" entries. Only applies to "time" assignments.
## Can be "utc", "local" or any valid Golang timezone (e.g. "Europe/Berlin")
entries = [
{ type = "string", assignment = "measurement", terminator = "null" },
{ name = "address", type = "uint16", assignment = "tag" },
{ name = "value", type = "float64" },
{ type = "unix", assignment = "time" },
]
## Optional: Filter evaluated before applying the configuration.
## This option can be used to mange multiple configuration specific for
## a certain message type. If no filter is given, the configuration is applied.
# [inputs.file.binary.filter]
# ## Filter message by the exact length in bytes (default: N/A).
# # length = 0
# ## Filter the message by a minimum length in bytes.
# ## Messages longer of equal length will pass.
# # length_min = 0
# ## List of data parts to match.
# ## Only if all selected parts match, the configuration will be
# ## applied. The "offset" is the start of the data to match in bits,
# ## "bits" is the length in bits and "match" is the value to match
# ## against. Non-byte boundaries are supported, data is always right-aligned.
# selection = [
# { offset = 0, bits = 8, match = "0x1F" },
# ]
#
#
```
In this configuration mode, you explicitly specify the field and tags you want
to scrape out of your data.
A configuration can contain multiple `binary` subsections for e.g. the file
plugin to process the binary data multiple times. This can be useful
(together with _filters_) to handle different message types.
__Please note__: The `filter` section needs to be placed __after__ the `entries`
definitions due to TOML constraints as otherwise the entries will be assigned
to the filter section.
### General options and remarks
#### `allow_no_match` (optional)
By specifying `allow_no_match` you allow the parser to silently ignore data
that does not match _any_ given configuration filter. This can be useful if
you only want to collect a subset of the available messages.
#### `endianness` (optional)
This specifies the endianness of the data. If not specified, the parser will
fallback to the "host" endianness, assuming that the message and Telegraf
machine share the same endianness.
Alternatively, you can explicitly specify big-endian format (`"be"`) or
little-endian format (`"le"`).
#### `binary_encoding` (optional)
If this option is not specified or set to `none`, the input data contains the
binary data as raw bytes. This is the default.
If set to `hex`, the input data is interpreted as a string containing
hex-encoded data like `C0 C7 21 A9`. The value is _case insensitive_ and can
handle spaces and prefixes like `0x` or `x`.
If set to `base64` the input data is interpreted as a string containing
padded base64 data `RDLAAA==`.
### Non-byte aligned value extraction
In both, `filter` and `entries` definitions, values can be extracted at non-byte
boundaries. You can for example extract 3-bit starting at bit-offset 8. In those
cases, the result will be masked and shifted such that the resulting byte-value
is _right_ aligned. In case your 3-bit are `101` the resulting byte value is
`0x05`.
This is especially important when specifying the `match` value in the filter
section.
### Entries definitions
The entry array specifies how to dissect the message into the measurement name,
the timestamp, tags and fields.
#### `measurement` specification
When setting the `assignment` to `"measurement"`, the extracted value
will be used as the metric name, overriding other specifications.
The `type` setting is assumed to be `"string"` and can be omitted similar
to the `name` option. See [`string` type handling](#string-type-handling)
for details and further options.
### `time` specification
When setting the `assignment` to `"time"`, the extracted value
will be used as the timestamp of the metric. By default the current
time will be used for all created metrics.
The `type` setting here contains the time-format can be set to `unix`,
`unix_ms`, `unix_us`, `unix_ns`, or an accepted
[Go "reference time"][time const]. Consult the Go [time][time parse]
package for details and additional examples on how to set the time format.
If `type` is omitted the `unix` format is assumed.
For the `unix` format and derivatives, the underlying value is assumed
to be a 64-bit integer. The `bits` setting can be used to specify other
length settings. All other time-formats assume a fixed-length `string`
value to be extracted. The length of the string is automatically
determined using the format setting in `type`.
The `timezone` setting allows to convert the extracted time to the
given value timezone. By default the time will be interpreted as `utc`.
Other valid values are `local`, i.e. the local timezone configured for
the machine, or valid timezone-specification e.g. `Europe/Berlin`.
### `tag` specification
When setting the `assignment` to `"tag"`, the extracted value
will be used as a tag. The `name` setting will be the name of the tag
and the `type` will default to `string`. When specifying other types,
the extracted value will first be interpreted as the given type and
then converted to `string`.
The `bits` setting can be used to specify the length of the data to
extract and is required for fixed-length `string` types.
### `field` specification
When setting the `assignment` to `"field"` or omitting the `assignment`
setting, the extracted value will be used as a field. The `name` setting
is used as the name of the field and the `type` as type of the field value.
The `bits` setting can be used to specify the length of the data to
extract. By default the length corresponding to `type` is used.
Please see the [string](#string-type-handling) and [bool](#bool-type-handling)
specific sections when using those types.
### `string` type handling
Strings are assumed to be fixed-length strings by default. In this case, the
`bits` setting is mandatory to specify the length of the string in _bit_.
To handle dynamic strings, the `terminator` setting can be used to specify
characters to terminate the string. The two named options, `fixed` and `null`
will specify fixed-length and null-terminated strings, respectively.
Any other setting will be interpreted as hexadecimal sequence of bytes
matching the end of the string. The termination-sequence is removed from
the result.
### `bool` type handling
By default `bool` types are assumed to be _one_ bit in length. You can
specify any other length by using the `bits` setting.
When interpreting values as booleans, any zero value will be `false`,
while any non-zero value will result in `true`.
### omitting data
Parts of the data can be omitted by setting `omit = true`. In this case,
you only need to specify the length of the chunk to omit by either using
the `type` or `bits` setting. All other options can be skipped.
### Filter definitions
Filters can be used to match the length or the content of the data against
a specified reference. See the [examples section](#examples) for details.
You can also check multiple parts of the message by specifying multiple
`section` entries for a filter. Each `section` is then matched separately.
All have to match to apply the configuration.
#### `length` and `length_min` options
Using the `length` option, the filter will check if the data to parse has
exactly the given number of _bytes_. Otherwise, the configuration will not
be applied.
Similarly, for `length_min` the data has to have _at least_ the given number
of _bytes_ to generate a match.
#### `selection` list
Selections can be used with or without length constraints to match the content
of the data. Here, the `offset` and `bits` properties will specify the start
and length of the data to check. Both values are in _bit_ allowing for non-byte
aligned value extraction. The extracted data will the be checked against the
given `match` value specified in HEX.
If multiple `selection` entries are specified _all_ of the selections must
match for the configuration to get applied.
## Examples
In the following example, we use a binary protocol with three different messages
in little-endian format
### Message A definition
```text
+--------+------+------+--------+--------+------------+--------------------+--------------------+
| ID | type | len | addr | count | failure | value | timestamp |
+--------+------+------+--------+--------+------------+--------------------+--------------------+
| 0x0201 | 0x0A | 0x18 | 0x7F01 | 0x2A00 | 0x00000000 | 0x6F1283C0CA210940 | 0x10D4DF6200000000 |
+--------+------+------+--------+--------+------------+--------------------+--------------------+
```
### Message B definition
```text
+--------+------+------+------------+
| ID | type | len | value |
+--------+------+------+------------+
| 0x0201 | 0x0B | 0x04 | 0xDEADC0DE |
+--------+------+------+------------+
```
### Message C definition
```text
+--------+------+------+------------+------------+--------------------+
| ID | type | len | value x | value y | timestamp |
+--------+------+------+------------+------------+--------------------+
| 0x0201 | 0x0C | 0x10 | 0x4DF82D40 | 0x5F305C08 | 0x10D4DF6200000000 |
+--------+------+------+------------+------------+--------------------+
```
All messages consists of a 4-byte header containing the _message type_
in the 3rd byte and a message specific body. To parse those messages
you can use the following configuration
```toml
[[inputs.file]]
files = ["messageA.bin", "messageB.bin", "messageC.bin"]
data_format = "binary"
endianness = "le"
[[inputs.file.binary]]
metric_name = "messageA"
entries = [
{ bits = 32, omit = true },
{ name = "address", type = "uint16", assignment = "tag" },
{ name = "count", type = "int16" },
{ name = "failure", type = "bool", bits = 32, assignment = "tag" },
{ name = "value", type = "float64" },
{ type = "unix", assignment = "time" },
]
[inputs.file.binary.filter]
selection = [{ offset = 16, bits = 8, match = "0x0A" }]
[[inputs.file.binary]]
metric_name = "messageB"
entries = [
{ bits = 32, omit = true },
{ name = "value", type = "uint32" },
]
[inputs.file.binary.filter]
selection = [{ offset = 16, bits = 8, match = "0x0B" }]
[[inputs.file.binary]]
metric_name = "messageC"
entries = [
{ bits = 32, omit = true },
{ name = "x", type = "float32" },
{ name = "y", type = "float32" },
{ type = "unix", assignment = "time" },
]
[inputs.file.binary.filter]
selection = [{ offset = 16, bits = 8, match = "0x0C" }]
```
The above configuration has one `[[inputs.file.binary]]` section per
message type and uses a filter in each of those sections to apply
the correct configuration by comparing the 3rd byte (containing
the message type). This will lead to the following output
```text
metricA,address=383,failure=false count=42i,value=3.1415 1658835984000000000
metricB value=3737169374i 1658847037000000000
metricC x=2.718280076980591,y=0.0000000000000000000000000000000006626070178575745 1658835984000000000
```
where `metricB` uses the parsing time as timestamp due to missing
information in the data. The other two metrics use the timestamp
derived from the data.
[time const]: https://golang.org/pkg/time/#pkg-constants
[time parse]: https://golang.org/pkg/time/#Parse

183
plugins/parsers/binary/config.go Normal file
View file

@ -0,0 +1,183 @@
package binary
import (
"encoding/binary"
"encoding/hex"
"errors"
"fmt"
"strings"
"time"
"github.com/influxdata/telegraf"
"github.com/influxdata/telegraf/internal"
"github.com/influxdata/telegraf/metric"
)
type BinaryPart struct {
Offset uint64 `toml:"offset"`
Bits uint64 `toml:"bits"`
Match string `toml:"match"`
val []byte
}
type Filter struct {
Selection []BinaryPart `toml:"selection"`
LengthMin uint64 `toml:"length_min"`
Length uint64 `toml:"length"`
}
type Config struct {
MetricName string `toml:"metric_name"`
Filter *Filter `toml:"filter"`
Entries []Entry `toml:"entries"`
}
func (c *Config) preprocess(defaultName string) error {
// Preprocess filter part
if c.Filter != nil {
if c.Filter.Length != 0 && c.Filter.LengthMin != 0 {
return errors.New("length and length_min cannot be used together")
}
var length uint64
for i, s := range c.Filter.Selection {
end := (s.Offset + s.Bits) / 8
if (s.Offset+s.Bits)%8 != 0 {
end++
}
if end > length {
length = end
}
var err error
s.val, err = hex.DecodeString(strings.TrimPrefix(s.Match, "0x"))
if err != nil {
return fmt.Errorf("decoding match %d failed: %w", i, err)
}
c.Filter.Selection[i] = s
}
if c.Filter.Length != 0 && length > c.Filter.Length {
return fmt.Errorf("filter length (%d) larger than constraint (%d)", length, c.Filter.Length)
}
if c.Filter.Length == 0 && length > c.Filter.LengthMin {
c.Filter.LengthMin = length
}
}
// Preprocess entries part
var hasField, hasMeasurement bool
defined := make(map[string]bool)
for i, e := range c.Entries {
if err := e.check(); err != nil {
return fmt.Errorf("entry %q (%d): %w", e.Name, i, err)
}
// Store the normalized entry
c.Entries[i] = e
if e.Omit {
continue
}
// Check for duplicate entries
key := e.Assignment + "_" + e.Name
if defined[key] {
return fmt.Errorf("multiple definitions of %q", e.Name)
}
defined[key] = true
hasMeasurement = hasMeasurement || e.Assignment == "measurement"
hasField = hasField || e.Assignment == "field"
}
if !hasMeasurement && c.MetricName == "" {
if defaultName == "" {
return errors.New("no metric name given")
}
c.MetricName = defaultName
}
if !hasField {
return errors.New("no field defined")
}
return nil
}
func (c *Config) matches(in []byte) bool {
// If no filter is given, just match everything
if c.Filter == nil {
return true
}
// Checking length constraints
length := uint64(len(in))
if c.Filter.Length != 0 && length != c.Filter.Length {
return false
}
if c.Filter.LengthMin != 0 && length < c.Filter.LengthMin {
return false
}
// Matching elements
for _, s := range c.Filter.Selection {
data, err := extractPart(in, s.Offset, s.Bits)
if err != nil {
return false
}
if len(data) != len(s.val) {
return false
}
for i, v := range data {
if v != s.val[i] {
return false
}
}
}
return true
}
func (c *Config) collect(in []byte, order binary.ByteOrder, defaultTime time.Time) (telegraf.Metric, error) {
t := defaultTime
name := c.MetricName
tags := make(map[string]string)
fields := make(map[string]interface{})
var offset uint64
for _, e := range c.Entries {
data, n, err := e.extract(in, offset)
if err != nil {
return nil, err
}
offset += n
switch e.Assignment {
case "measurement":
name = convertStringType(data)
case "field":
v, err := e.convertType(data, order)
if err != nil {
return nil, fmt.Errorf("field %q failed: %w", e.Name, err)
}
fields[e.Name] = v
case "tag":
raw, err := e.convertType(data, order)
if err != nil {
return nil, fmt.Errorf("tag %q failed: %w", e.Name, err)
}
v, err := internal.ToString(raw)
if err != nil {
return nil, fmt.Errorf("tag %q failed: %w", e.Name, err)
}
tags[e.Name] = v
case "time":
var err error
t, err = e.convertTimeType(data, order)
if err != nil {
return nil, fmt.Errorf("time failed: %w", err)
}
}
}
return metric.New(name, tags, fields, t), nil
}

310
plugins/parsers/binary/entry.go Normal file
View file

@ -0,0 +1,310 @@
package binary
import (
"encoding/binary"
"encoding/hex"
"errors"
"fmt"
"math"
"strings"
"time"
"github.com/influxdata/telegraf/internal"
)
type Entry struct {
Name string `toml:"name"`
Type string `toml:"type"`
Bits uint64 `toml:"bits"`
Omit bool `toml:"omit"`
Terminator string `toml:"terminator"`
Timezone string `toml:"timezone"`
Assignment string `toml:"assignment"`
termination []byte
location *time.Location
}
func (e *Entry) check() error {
// Normalize cases
e.Assignment = strings.ToLower(e.Assignment)
e.Terminator = strings.ToLower(e.Terminator)
if e.Assignment != "time" {
e.Type = strings.ToLower(e.Type)
}
// Handle omitted fields
if e.Omit {
if e.Bits == 0 && e.Type == "" {
return errors.New("neither type nor bits given")
}
if e.Bits == 0 {
bits, err := bitsForType(e.Type)
if err != nil {
return err
}
e.Bits = bits
}
return nil
}
// Set name for global options
if e.Assignment == "measurement" || e.Assignment == "time" {
e.Name = e.Assignment
}
// Check the name
if e.Name == "" {
return errors.New("missing name")
}
// Check the assignment
var defaultType string
switch e.Assignment {
case "measurement":
defaultType = "string"
if e.Type != "string" && e.Type != "" {
return errors.New("'measurement' type has to be 'string'")
}
case "time":
bits := uint64(64)
switch e.Type {
// Make 'unix' the default
case "":
defaultType = "unix"
// Special plugin specific names
case "unix", "unix_ms", "unix_us", "unix_ns":
// Format-specification string formats
default:
bits = uint64(len(e.Type) * 8)
}
if e.Bits == 0 {
e.Bits = bits
}
switch e.Timezone {
case "", "utc":
// Make UTC the default
e.location = time.UTC
case "local":
e.location = time.Local
default:
var err error
e.location, err = time.LoadLocation(e.Timezone)
if err != nil {
return err
}
}
case "tag":
defaultType = "string"
case "", "field":
e.Assignment = "field"
default:
return fmt.Errorf("no assignment for %q", e.Name)
}
// Check type (special type for "time")
switch e.Type {
case "uint8", "int8", "uint16", "int16", "uint32", "int32", "uint64", "int64":
fallthrough
case "float32", "float64":
bits, err := bitsForType(e.Type)
if err != nil {
return err
}
if e.Bits == 0 {
e.Bits = bits
}
if bits < e.Bits {
return fmt.Errorf("type overflow for %q", e.Name)
}
case "bool":
if e.Bits == 0 {
e.Bits = 1
}
case "string":
// Check termination
switch e.Terminator {
case "", "fixed":
e.Terminator = "fixed"
if e.Bits == 0 {
return fmt.Errorf("require 'bits' for fixed-length string for %q", e.Name)
}
case "null":
e.termination = []byte{0}
if e.Bits != 0 {
return fmt.Errorf("cannot use 'bits' and 'null' terminator together for %q", e.Name)
}
default:
if e.Bits != 0 {
return fmt.Errorf("cannot use 'bits' and terminator together for %q", e.Name)
}
var err error
e.termination, err = hex.DecodeString(strings.TrimPrefix(e.Terminator, "0x"))
if err != nil {
return fmt.Errorf("decoding terminator failed for %q: %w", e.Name, err)
}
}
// We can only handle strings that adhere to byte-bounds
if e.Bits%8 != 0 {
return fmt.Errorf("non-byte length for string field %q", e.Name)
}
case "":
if defaultType == "" {
return fmt.Errorf("no type for %q", e.Name)
}
e.Type = defaultType
default:
if e.Assignment != "time" {
return fmt.Errorf("unknown type for %q", e.Name)
}
}
return nil
}
func (e *Entry) extract(in []byte, offset uint64) ([]byte, uint64, error) {
if e.Bits > 0 {
data, err := extractPart(in, offset, e.Bits)
return data, e.Bits, err
}
if e.Type != "string" {
return nil, 0, fmt.Errorf("unexpected entry: %v", e)
}
inbits := uint64(len(in)) * 8
// Read up to the termination
var found bool
var data []byte
var termOffset int
var n uint64
for offset+n+8 <= inbits {
buf, err := extractPart(in, offset+n, 8)
if err != nil {
return nil, 0, err
}
if len(buf) != 1 {
return nil, 0, fmt.Errorf("unexpected length %d", len(buf))
}
data = append(data, buf[0])
n += 8
// Check for terminator
if buf[0] == e.termination[termOffset] {
termOffset++
}
if termOffset == len(e.termination) {
found = true
break
}
}
if !found {
return nil, n, fmt.Errorf("terminator not found for %q", e.Name)
}
// Strip the terminator
return data[:len(data)-len(e.termination)], n, nil
}
func (e *Entry) convertType(in []byte, order binary.ByteOrder) (interface{}, error) {
switch e.Type {
case "uint8", "int8", "uint16", "int16", "uint32", "int32", "float32", "uint64", "int64", "float64":
return convertNumericType(in, e.Type, order)
case "bool":
return convertBoolType(in), nil
case "string":
return convertStringType(in), nil
}
return nil, fmt.Errorf("cannot handle type %q", e.Type)
}
func (e *Entry) convertTimeType(in []byte, order binary.ByteOrder) (time.Time, error) {
factor := int64(1)
switch e.Type {
case "unix":
factor *= 1000
fallthrough
case "unix_ms":
factor *= 1000
fallthrough
case "unix_us":
factor *= 1000
fallthrough
case "unix_ns":
raw, err := convertNumericType(in, "int64", order)
if err != nil {
return time.Unix(0, 0), err
}
v := raw.(int64)
return time.Unix(0, v*factor).In(e.location), nil
}
// We have a format specification (hopefully)
v := convertStringType(in)
return internal.ParseTimestamp(e.Type, v, e.location)
}
func convertStringType(in []byte) string {
return string(in)
}
func convertNumericType(in []byte, t string, order binary.ByteOrder) (interface{}, error) {
bits, err := bitsForType(t)
if err != nil {
return nil, err
}
inlen := uint64(len(in))
expected := bits / 8
if inlen > expected {
// Should never happen
return 0, fmt.Errorf("too many bytes %d vs %d", len(in), expected)
}
// Pad the data if shorter than the datatype length
buf := make([]byte, expected-inlen, expected)
buf = append(buf, in...)
switch t {
case "uint8":
return buf[0], nil
case "int8":
return int8(buf[0]), nil
case "uint16":
return order.Uint16(buf), nil
case "int16":
v := order.Uint16(buf)
return int16(v), nil
case "uint32":
return order.Uint32(buf), nil
case "int32":
v := order.Uint32(buf)
return int32(v), nil
case "uint64":
return order.Uint64(buf), nil
case "int64":
v := order.Uint64(buf)
return int64(v), nil
case "float32":
v := order.Uint32(buf)
return math.Float32frombits(v), nil
case "float64":
v := order.Uint64(buf)
return math.Float64frombits(v), nil
}
return nil, fmt.Errorf("no numeric type %q", t)
}
func convertBoolType(in []byte) bool {
for _, x := range in {
if x != 0 {
return true
}
}
return false
}

44
plugins/parsers/binary/entry_test.go Normal file
View file

@ -0,0 +1,44 @@
package binary
import (
"testing"
"time"
"github.com/stretchr/testify/require"
"github.com/influxdata/telegraf/internal"
)
func TestEntryExtract(t *testing.T) {
testdata := []byte{0x01, 0x02, 0x03, 0x04}
e := &Entry{Type: "uint64"}
_, _, err := e.extract(testdata, 0)
require.EqualError(t, err, `unexpected entry: &{ uint64 0 false [] <nil>}`)
}
func TestEntryConvertType(t *testing.T) {
testdata := []byte{0x01, 0x02, 0x03, 0x04}
e := &Entry{Type: "garbage"}
_, err := e.convertType(testdata, internal.HostEndianness)
require.EqualError(t, err, `cannot handle type "garbage"`)
}
func TestEntryConvertTimeType(t *testing.T) {
testdata := []byte{0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09}
e := &Entry{Type: "unix_ns", location: time.UTC}
_, err := e.convertTimeType(testdata, internal.HostEndianness)
require.EqualError(t, err, `too many bytes 9 vs 8`)
}
func TestConvertNumericType(t *testing.T) {
testdata := []byte{0x01, 0x02, 0x03, 0x04}
_, err := convertNumericType(testdata, "garbage", internal.HostEndianness)
require.EqualError(t, err, `cannot determine length for type "garbage"`)
_, err = convertNumericType(testdata, "uint8", internal.HostEndianness)
require.EqualError(t, err, `too many bytes 4 vs 1`)
}

213
plugins/parsers/binary/parser.go Normal file
View file

@ -0,0 +1,213 @@
package binary
import (
"encoding/base64"
"encoding/binary"
"encoding/hex"
"errors"
"fmt"
"strings"
"time"
"github.com/influxdata/telegraf"
"github.com/influxdata/telegraf/internal"
"github.com/influxdata/telegraf/plugins/parsers"
)
type Parser struct {
AllowNoMatch bool `toml:"allow_no_match"`
Endianess string `toml:"endianess" deprecated:"1.27.4;1.35.0;use 'endianness' instead"`
Endianness string `toml:"endianness"`
Configs []Config `toml:"binary"`
HexEncoding bool `toml:"hex_encoding" deprecated:"1.30.0;1.35.0;use 'binary_encoding' instead"`
Encoding string `toml:"binary_encoding"`
Log telegraf.Logger `toml:"-"`
metricName string
defaultTags map[string]string
converter binary.ByteOrder
}
func (p *Parser) Init() error {
// Keep backward compatibility
if p.Endianess != "" && p.Endianness == "" {
p.Endianness = p.Endianess
}
if p.HexEncoding {
if p.Encoding != "" && p.Encoding != "hex" {
return errors.New("conflicting settings between 'hex_encoding' and 'binary_encoding'")
}
p.Encoding = "hex"
}
switch p.Endianness {
case "le":
p.converter = binary.LittleEndian
case "be":
p.converter = binary.BigEndian
case "", "host":
p.converter = internal.HostEndianness
default:
return fmt.Errorf("unknown endianness %q", p.Endianness)
}
switch p.Encoding {
case "", "none", "hex", "base64":
default:
return fmt.Errorf("unknown encoding %q", p.Encoding)
}
// Pre-process the configurations
if len(p.Configs) == 0 {
return errors.New("no configuration given")
}
for i, cfg := range p.Configs {
if err := cfg.preprocess(p.metricName); err != nil {
return fmt.Errorf("config %d invalid: %w", i, err)
}
p.Configs[i] = cfg
}
return nil
}
func (p *Parser) Parse(data []byte) ([]telegraf.Metric, error) {
t := time.Now()
// If the data is encoded in HEX, we need to decode it first
buf := data
switch p.Encoding {
case "hex":
s := strings.TrimPrefix(string(data), "0x")
s = strings.TrimPrefix(s, "x")
s = strings.TrimSpace(s)
s = strings.ReplaceAll(s, " ", "")
s = strings.ReplaceAll(s, "\t", "")
var err error
buf, err = hex.DecodeString(s)
if err != nil {
return nil, fmt.Errorf("decoding hex failed: %w", err)
}
case "base64":
decoder := base64.StdEncoding.WithPadding(base64.StdPadding)
var err error
buf, err = decoder.DecodeString(strings.TrimSpace(string(data)))
if err != nil {
return nil, fmt.Errorf("decoding base64 failed: %w", err)
}
}
matches := 0
metrics := make([]telegraf.Metric, 0)
for i, cfg := range p.Configs {
// Apply the filter and see if we should match this
if !cfg.matches(buf) {
p.Log.Debugf("ignoring data in config %d", i)
continue
}
matches++
// Collect the metric
m, err := cfg.collect(buf, p.converter, t)
if err != nil {
return nil, err
}
metrics = append(metrics, m)
}
if matches == 0 && !p.AllowNoMatch {
return nil, errors.New("no matching configuration")
}
return metrics, nil
}
func (p *Parser) ParseLine(line string) (telegraf.Metric, error) {
metrics, err := p.Parse([]byte(line))
if err != nil {
return nil, err
}
switch len(metrics) {
case 0:
return nil, nil
case 1:
return metrics[0], nil
default:
return metrics[0], fmt.Errorf("cannot parse line with multiple (%d) metrics", len(metrics))
}
}
func (p *Parser) SetDefaultTags(tags map[string]string) {
p.defaultTags = tags
}
func init() {
// Register all variants
parsers.Add("binary",
func(defaultMetricName string) telegraf.Parser {
return &Parser{metricName: defaultMetricName}
},
)
}
func extractPart(in []byte, offset, bits uint64) ([]byte, error) {
inLen := uint64(len(in))
start := offset / 8
bitend := offset%8 + bits
length := bitend / 8
if bitend%8 != 0 {
length++
}
if start+length > inLen {
return nil, fmt.Errorf("out-of-bounds @%d with %d bits", offset, bits)
}
var out []byte
out = append(out, in[start:start+length]...)
if offset%8 != 0 {
// Mask the start-byte with the non-aligned bit-mask
startmask := (byte(1) << (8 - offset%8)) - 1
out[0] = out[0] & startmask
}
if bitend%8 == 0 {
// The end is aligned to byte-boundaries
return out, nil
}
shift := 8 - bitend%8
carryshift := bitend % 8
// We need to shift right in case of not ending at a byte boundary
// to make the bits right aligned.
// Carry over the bits from the byte left to fill in...
var carry byte
for i, x := range out {
out[i] = (x >> shift) | carry
carry = x << carryshift
}
if bits%8 == 0 {
// Avoid an empty leading byte
return out[1:], nil
}
return out, nil
}
func bitsForType(t string) (uint64, error) {
switch t {
case "uint8", "int8":
return 8, nil
case "uint16", "int16":
return 16, nil
case "uint32", "int32", "float32":
return 32, nil
case "uint64", "int64", "float64":
return 64, nil
}
return 0, fmt.Errorf("cannot determine length for type %q", t)
}

1614
plugins/parsers/binary/parser_test.go Normal file
View file

File diff suppressed because it is too large Load diff

3
plugins/parsers/binary/testcases/base64_encoding/expected.out Normal file
View file

@ -0,0 +1,3 @@
test value=715
test value=208.5
test value=0.471

1
plugins/parsers/binary/testcases/base64_encoding/messageA.bin Normal file
View file

@ -0,0 +1 @@
RDLAAA==

1
plugins/parsers/binary/testcases/base64_encoding/messageB.bin Normal file
View file

@ -0,0 +1 @@
Q1CAAA==

1
plugins/parsers/binary/testcases/base64_encoding/messageC.bin Normal file
View file

@ -0,0 +1 @@
PvEm6Q==

8
plugins/parsers/binary/testcases/base64_encoding/telegraf.conf Normal file
View file

@ -0,0 +1,8 @@
[[inputs.test]]
files = ["messageA.bin", "messageB.bin", "messageC.bin"]
data_format = "binary"
endianness = "be"
binary_encoding = "base64"
[[inputs.test.binary]]
entries = [{ name = "value", type = "float32" }]

3
plugins/parsers/binary/testcases/hex_encoding/expected.out Normal file
View file

@ -0,0 +1,3 @@
test value=715
test value=208.5
test value=0.471

1
plugins/parsers/binary/testcases/hex_encoding/messageA.bin Normal file
View file

@ -0,0 +1 @@
0x4432c000

1
plugins/parsers/binary/testcases/hex_encoding/messageB.bin Normal file
View file

@ -0,0 +1 @@
0x43508000

1
plugins/parsers/binary/testcases/hex_encoding/messageC.bin Normal file
View file

@ -0,0 +1 @@
0x3ef126e9

8
plugins/parsers/binary/testcases/hex_encoding/telegraf.conf Normal file
View file

@ -0,0 +1,8 @@
[[inputs.test]]
files = ["messageA.bin", "messageB.bin", "messageC.bin"]
data_format = "binary"
endianness = "be"
binary_encoding = "hex"
[[inputs.test.binary]]
entries = [{ name = "value", type = "float32" }]

3
plugins/parsers/binary/testcases/multiple_messages/expected.out Normal file
View file

@ -0,0 +1,3 @@
metricA,address=383,failure=false count=42i,value=3.1415 1658835984000000000
metricB value=3737169374u 1658835984000000000
metricC x=2.718280076980591,y=0.0000000000000000000000000000000006626070178575745 1658835984000000000

BIN
plugins/parsers/binary/testcases/multiple_messages/messageA.bin Normal file
View file

Binary file not shown.

BIN
plugins/parsers/binary/testcases/multiple_messages/messageB.bin Normal file
View file

Binary file not shown.

BIN
plugins/parsers/binary/testcases/multiple_messages/messageC.bin Normal file
View file

Binary file not shown.

46
plugins/parsers/binary/testcases/multiple_messages/telegraf.conf Normal file
View file

@ -0,0 +1,46 @@
[[inputs.test]]
files = ["messageA.bin", "messageB.bin", "messageC.bin"]
data_format = "binary"
endianness = "le"
[[inputs.test.binary]]
metric_name = "metricA"
entries = [
{ bits = 32, omit = true },
{ name = "address", type = "uint16", assignment = "tag" },
{ name = "count", type = "int16" },
{ name = "failure", type = "bool", bits = 32, assignment = "tag" },
{ name = "value", type = "float64" },
{ type = "unix", assignment = "time" },
]
[inputs.test.binary.filter]
selection = [
{ offset = 16, bits = 8, match = "0x0A" },
]
[[inputs.test.binary]]
metric_name = "metricB"
entries = [
{ bits = 32, omit = true },
{ name = "value", type = "uint32" },
{ type = "unix", assignment = "time" },
]
[inputs.test.binary.filter]
selection = [{ offset = 16, bits = 8, match = "0x0B" }]
[[inputs.test.binary]]
metric_name = "metricC"
entries = [
{ bits = 32, omit = true },
{ name = "x", type = "float32" },
{ name = "y", type = "float32" },
{ type = "unix", assignment = "time" },
]
[inputs.test.binary.filter]
selection = [{ offset = 16, bits = 8, match = "0x0C" }]