441 lines
12 KiB
Go
441 lines
12 KiB
Go
package internal
|
|
|
|
import (
|
|
"bufio"
|
|
"compress/gzip"
|
|
"context"
|
|
crypto_rand "crypto/rand"
|
|
"errors"
|
|
"fmt"
|
|
"io"
|
|
"log"
|
|
"math/big"
|
|
"math/rand"
|
|
"os"
|
|
"os/exec"
|
|
"runtime"
|
|
"strings"
|
|
"sync"
|
|
"syscall"
|
|
"time"
|
|
"unicode"
|
|
|
|
"github.com/influxdata/telegraf/internal/choice"
|
|
)
|
|
|
|
const alphanum string = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz"
|
|
const NoMetricsCreatedMsg = "No metrics were created from a message. Verify your parser settings. This message is only printed once."
|
|
|
|
var once sync.Once
|
|
|
|
var (
|
|
ErrTimeout = errors.New("command timed out")
|
|
ErrNotImplemented = errors.New("not implemented yet")
|
|
)
|
|
|
|
// Set via LDFLAGS -X
|
|
var (
|
|
Version = "unknown"
|
|
Branch = ""
|
|
Commit = ""
|
|
)
|
|
|
|
type ReadWaitCloser struct {
|
|
pipeReader *io.PipeReader
|
|
wg sync.WaitGroup
|
|
}
|
|
|
|
func FormatFullVersion() string {
|
|
var parts = []string{"Telegraf"}
|
|
|
|
if Version != "" {
|
|
parts = append(parts, Version)
|
|
} else {
|
|
parts = append(parts, "unknown")
|
|
}
|
|
|
|
if Branch != "" || Commit != "" {
|
|
if Branch == "" {
|
|
Branch = "unknown"
|
|
}
|
|
if Commit == "" {
|
|
Commit = "unknown"
|
|
}
|
|
git := fmt.Sprintf("(git: %s@%s)", Branch, Commit)
|
|
parts = append(parts, git)
|
|
}
|
|
|
|
return strings.Join(parts, " ")
|
|
}
|
|
|
|
// ProductToken returns a tag for Telegraf that can be used in user agents.
|
|
func ProductToken() string {
|
|
return fmt.Sprintf("Telegraf/%s Go/%s",
|
|
Version, strings.TrimPrefix(runtime.Version(), "go"))
|
|
}
|
|
|
|
// ReadLines reads contents from a file and splits them by new lines.
|
|
func ReadLines(filename string) ([]string, error) {
|
|
f, err := os.Open(filename)
|
|
if err != nil {
|
|
return []string{""}, err
|
|
}
|
|
defer f.Close()
|
|
|
|
var ret []string
|
|
scanner := bufio.NewScanner(f)
|
|
for scanner.Scan() {
|
|
ret = append(ret, scanner.Text())
|
|
}
|
|
|
|
return ret, nil
|
|
}
|
|
|
|
// RandomString returns a random string of alphanumeric characters
|
|
func RandomString(n int) (string, error) {
|
|
var bytes = make([]byte, n)
|
|
_, err := crypto_rand.Read(bytes)
|
|
if err != nil {
|
|
return "", err
|
|
}
|
|
for i, b := range bytes {
|
|
bytes[i] = alphanum[b%byte(len(alphanum))]
|
|
}
|
|
return string(bytes), nil
|
|
}
|
|
|
|
// SnakeCase converts the given string to snake case following the Golang format:
|
|
// acronyms are converted to lower-case and preceded by an underscore.
|
|
func SnakeCase(in string) string {
|
|
runes := []rune(in)
|
|
length := len(runes)
|
|
|
|
var out []rune
|
|
for i := 0; i < length; i++ {
|
|
if i > 0 && unicode.IsUpper(runes[i]) {
|
|
prevLower := unicode.IsLower(runes[i-1])
|
|
nextLower := i+1 < length && unicode.IsLower(runes[i+1])
|
|
// Special case for plural acronyms
|
|
nextPlural := i+1 < length && runes[i+1] == 's'
|
|
|
|
if prevLower || (nextLower && !nextPlural) {
|
|
out = append(out, '_')
|
|
}
|
|
}
|
|
out = append(out, unicode.ToLower(runes[i]))
|
|
}
|
|
|
|
return string(out)
|
|
}
|
|
|
|
// RandomSleep will sleep for a random amount of time up to max.
|
|
// If the shutdown channel is closed, it will return before it has finished sleeping.
|
|
func RandomSleep(limit time.Duration, shutdown chan struct{}) {
|
|
sleepDuration := RandomDuration(limit)
|
|
if sleepDuration == 0 {
|
|
return
|
|
}
|
|
|
|
t := time.NewTimer(time.Nanosecond * sleepDuration)
|
|
select {
|
|
case <-t.C:
|
|
return
|
|
case <-shutdown:
|
|
t.Stop()
|
|
return
|
|
}
|
|
}
|
|
|
|
// RandomDuration returns a random duration between 0 and max.
|
|
func RandomDuration(limit time.Duration) time.Duration {
|
|
if limit == 0 {
|
|
return 0
|
|
}
|
|
|
|
return time.Duration(rand.Int63n(limit.Nanoseconds())) //nolint:gosec // G404: not security critical
|
|
}
|
|
|
|
// SleepContext sleeps until the context is closed or the duration is reached.
|
|
func SleepContext(ctx context.Context, duration time.Duration) error {
|
|
if duration == 0 {
|
|
return nil
|
|
}
|
|
|
|
t := time.NewTimer(duration)
|
|
select {
|
|
case <-t.C:
|
|
return nil
|
|
case <-ctx.Done():
|
|
t.Stop()
|
|
return ctx.Err()
|
|
}
|
|
}
|
|
|
|
// AlignDuration returns the duration until next aligned interval.
|
|
// If the current time is aligned a 0 duration is returned.
|
|
func AlignDuration(tm time.Time, interval time.Duration) time.Duration {
|
|
return AlignTime(tm, interval).Sub(tm)
|
|
}
|
|
|
|
// AlignTime returns the time of the next aligned interval.
|
|
// If the current time is aligned the current time is returned.
|
|
func AlignTime(tm time.Time, interval time.Duration) time.Time {
|
|
truncated := tm.Truncate(interval)
|
|
if truncated.Equal(tm) {
|
|
return tm
|
|
}
|
|
return truncated.Add(interval)
|
|
}
|
|
|
|
// ExitStatus takes the error from exec.Command
|
|
// and returns the exit status and true
|
|
// if error is not exit status, will return 0 and false
|
|
func ExitStatus(err error) (int, bool) {
|
|
var exitErr *exec.ExitError
|
|
if errors.As(err, &exitErr) {
|
|
if status, ok := exitErr.Sys().(syscall.WaitStatus); ok {
|
|
return status.ExitStatus(), true
|
|
}
|
|
}
|
|
return 0, false
|
|
}
|
|
|
|
func (r *ReadWaitCloser) Close() error {
|
|
err := r.pipeReader.Close()
|
|
r.wg.Wait() // wait for the gzip goroutine finish
|
|
return err
|
|
}
|
|
|
|
// CompressWithGzip takes an io.Reader as input and pipes it through a
|
|
// gzip.Writer returning an io.Reader containing the gzipped data.
|
|
// Errors occurring during compression are returned to the instance reading
|
|
// from the returned reader via through the corresponding read call
|
|
// (e.g. io.Copy or io.ReadAll).
|
|
func CompressWithGzip(data io.Reader) io.ReadCloser {
|
|
pipeReader, pipeWriter := io.Pipe()
|
|
gzipWriter := gzip.NewWriter(pipeWriter)
|
|
|
|
// Start copying from the uncompressed reader to the output reader
|
|
// in the background until the input reader is closed (or errors out).
|
|
go func() {
|
|
// This copy will block until "data" reached EOF or an error occurs
|
|
_, err := io.Copy(gzipWriter, data)
|
|
|
|
// Close the compression writer and make sure we do not overwrite
|
|
// the copy error if any.
|
|
gzipErr := gzipWriter.Close()
|
|
if err == nil {
|
|
err = gzipErr
|
|
}
|
|
|
|
// Subsequent reads from the output reader (connected to "pipeWriter"
|
|
// via pipe) will return the copy (or closing) error if any to the
|
|
// instance reading from the reader returned by the CompressWithGzip
|
|
// function. If "err" is nil, the below function will correctly report
|
|
// io.EOF.
|
|
pipeWriter.CloseWithError(err)
|
|
}()
|
|
|
|
// Return a reader which then can be read by the caller to collect the
|
|
// compressed stream.
|
|
return pipeReader
|
|
}
|
|
|
|
// ParseTimestamp parses a Time according to the standard Telegraf options.
|
|
// These are generally displayed in the toml similar to:
|
|
//
|
|
// json_time_key= "timestamp"
|
|
// json_time_format = "2006-01-02T15:04:05Z07:00"
|
|
// json_timezone = "America/Los_Angeles"
|
|
//
|
|
// The format can be one of "unix", "unix_ms", "unix_us", "unix_ns", or a Go
|
|
// time layout suitable for time.Parse.
|
|
//
|
|
// When using the "unix" format, an optional fractional component is allowed.
|
|
// Specific unix time precisions cannot have a fractional component.
|
|
//
|
|
// Unix times may be an int64, float64, or string. When using a Go format
|
|
// string the timestamp must be a string.
|
|
//
|
|
// The location is a location string suitable for time.LoadLocation. Unix
|
|
// times do not use the location string, a unix time is always return in the
|
|
// UTC location.
|
|
func ParseTimestamp(format string, timestamp interface{}, location *time.Location, separator ...string) (time.Time, error) {
|
|
switch format {
|
|
case "unix", "unix_ms", "unix_us", "unix_ns":
|
|
sep := []string{",", "."}
|
|
if len(separator) > 0 {
|
|
sep = separator
|
|
}
|
|
return parseUnix(format, timestamp, sep)
|
|
default:
|
|
v, ok := timestamp.(string)
|
|
if !ok {
|
|
return time.Unix(0, 0), errors.New("unsupported type")
|
|
}
|
|
return parseTime(format, v, location)
|
|
}
|
|
}
|
|
|
|
// parseTime parses a timestamp in unix format with different resolutions
|
|
func parseUnix(format string, timestamp interface{}, separator []string) (time.Time, error) {
|
|
// Extract the scaling factor to nanoseconds from "format"
|
|
var factor int64
|
|
switch format {
|
|
case "unix":
|
|
factor = int64(time.Second)
|
|
case "unix_ms":
|
|
factor = int64(time.Millisecond)
|
|
case "unix_us":
|
|
factor = int64(time.Microsecond)
|
|
case "unix_ns":
|
|
factor = int64(time.Nanosecond)
|
|
}
|
|
|
|
zero := time.Unix(0, 0)
|
|
|
|
// Convert the representation to time
|
|
switch v := timestamp.(type) {
|
|
case int, int8, int16, int32, int64, uint, uint8, uint16, uint32, uint64:
|
|
t, err := ToInt64(v)
|
|
if err != nil {
|
|
return zero, err
|
|
}
|
|
return time.Unix(0, t*factor).UTC(), nil
|
|
case float32, float64:
|
|
ts, err := ToFloat64(v)
|
|
if err != nil {
|
|
return zero, err
|
|
}
|
|
|
|
// Parse the float as a precise fraction to avoid precision loss
|
|
f := big.Rat{}
|
|
if f.SetFloat64(ts) == nil {
|
|
return zero, errors.New("invalid number")
|
|
}
|
|
return timeFromFraction(&f, factor), nil
|
|
case string:
|
|
// Sanitize the string to have no thousand separators and dot
|
|
// as decimal separator to ease later parsing
|
|
v = sanitizeTimestamp(v, separator)
|
|
|
|
// Parse the string as a precise fraction to avoid precision loss
|
|
f := big.Rat{}
|
|
if _, ok := f.SetString(v); !ok {
|
|
return zero, errors.New("invalid number")
|
|
}
|
|
return timeFromFraction(&f, factor), nil
|
|
}
|
|
|
|
return zero, errors.New("unsupported type")
|
|
}
|
|
|
|
func timeFromFraction(f *big.Rat, factor int64) time.Time {
|
|
// Extract the numerator and denominator and scale to nanoseconds
|
|
num := f.Num()
|
|
denom := f.Denom()
|
|
num.Mul(num, big.NewInt(factor))
|
|
|
|
// Get the integer (non-fractional part) of the timestamp and convert
|
|
// it into time
|
|
t := big.Int{}
|
|
t.Div(num, denom)
|
|
|
|
return time.Unix(0, t.Int64()).UTC()
|
|
}
|
|
|
|
// sanitizeTimestamp removes thousand separators and uses dot as
|
|
// decimal separator. Returns also a boolean indicating success.
|
|
func sanitizeTimestamp(timestamp string, decimalSeparator []string) string {
|
|
// Remove thousand-separators that are not used for decimal separation
|
|
sanitized := timestamp
|
|
for _, s := range []string{" ", ",", "."} {
|
|
if !choice.Contains(s, decimalSeparator) {
|
|
sanitized = strings.ReplaceAll(sanitized, s, "")
|
|
}
|
|
}
|
|
|
|
// Replace decimal separators by dot to have a standard, parsable format
|
|
for _, s := range decimalSeparator {
|
|
// Make sure we replace only the first occurrence of any separator.
|
|
if strings.Contains(sanitized, s) {
|
|
return strings.Replace(sanitized, s, ".", 1)
|
|
}
|
|
}
|
|
return sanitized
|
|
}
|
|
|
|
// parseTime parses a string timestamp according to the format string.
|
|
func parseTime(format, timestamp string, location *time.Location) (time.Time, error) {
|
|
loc := location
|
|
if loc == nil {
|
|
loc = time.UTC
|
|
}
|
|
|
|
switch strings.ToLower(format) {
|
|
case "ansic":
|
|
format = time.ANSIC
|
|
case "unixdate":
|
|
format = time.UnixDate
|
|
case "rubydate":
|
|
format = time.RubyDate
|
|
case "rfc822":
|
|
format = time.RFC822
|
|
case "rfc822z":
|
|
format = time.RFC822Z
|
|
case "rfc850":
|
|
format = time.RFC850
|
|
case "rfc1123":
|
|
format = time.RFC1123
|
|
case "rfc1123z":
|
|
format = time.RFC1123Z
|
|
case "rfc3339":
|
|
format = time.RFC3339
|
|
case "rfc3339nano":
|
|
format = time.RFC3339Nano
|
|
case "stamp":
|
|
format = time.Stamp
|
|
case "stampmilli":
|
|
format = time.StampMilli
|
|
case "stampmicro":
|
|
format = time.StampMicro
|
|
case "stampnano":
|
|
format = time.StampNano
|
|
}
|
|
|
|
if !strings.Contains(format, "MST") {
|
|
return time.ParseInLocation(format, timestamp, loc)
|
|
}
|
|
|
|
// Golang does not parse times with ambiguous timezone abbreviations,
|
|
// but only parses the time-fields and the timezone NAME with a zero
|
|
// offset (see https://groups.google.com/g/golang-nuts/c/hDMdnm_jUFQ/m/yeL9IHOsAQAJ).
|
|
// To handle those timezones correctly we can use the timezone-name and
|
|
// force parsing the time in that timezone. This way we get the correct
|
|
// time for the "most probably" of the ambiguous timezone-abbreviations.
|
|
ts, err := time.Parse(format, timestamp)
|
|
if err != nil {
|
|
return time.Time{}, err
|
|
}
|
|
zone, offset := ts.Zone()
|
|
if zone == "UTC" || offset != 0 {
|
|
return ts.In(loc), nil
|
|
}
|
|
once.Do(func() {
|
|
const msg = `Your config is using abbreviated timezones and parsing was changed in v1.27.0!
|
|
Please see the change log, remove any workarounds in place, and carefully
|
|
check your data timestamps! If case you experience any problems, please
|
|
file an issue on https://github.com/influxdata/telegraf/issues!`
|
|
log.Print("W! " + msg)
|
|
})
|
|
|
|
abbrevLoc, err := time.LoadLocation(zone)
|
|
if err != nil {
|
|
return time.Time{}, fmt.Errorf("cannot resolve timezone abbreviation %q: %w", zone, err)
|
|
}
|
|
ts, err = time.ParseInLocation(format, timestamp, abbrevLoc)
|
|
if err != nil {
|
|
return time.Time{}, err
|
|
}
|
|
return ts.In(loc), nil
|
|
}
|