app/vlinsert: accept unix timestamp in seconds additionally to milliseconds at ElasticSearch bulk API

Timestamps in seconds are sometimes used for data ingestion via ElasticSearch bulk API
This commit is contained in:
Aliaksandr Valialkin 2024-09-28 21:17:10 +02:00
parent fd890a2771
commit 7d7d7c03bc
No known key found for this signature in database
GPG key ID: 52C003EE2BCDB9EB
4 changed files with 23 additions and 6 deletions

View file

@ -251,11 +251,15 @@ func parseElasticsearchTimestamp(s string) (int64, error) {
return 0, nil
}
if len(s) < len("YYYY-MM-DD") || s[len("YYYY")] != '-' {
// Try parsing timestamp in milliseconds
// Try parsing timestamp in seconds or milliseconds
n, err := strconv.ParseInt(s, 10, 64)
if err != nil {
return 0, fmt.Errorf("cannot parse timestamp in milliseconds from %q: %w", s, err)
}
if n < (1<<31) && n >= (-1<<31) {
// The timestamp is in seconds. Convert it to milliseconds
n *= 1e3
}
if n > int64(math.MaxInt64)/1e6 {
return 0, fmt.Errorf("too big timestamp in milliseconds: %d; mustn't exceed %d", n, int64(math.MaxInt64)/1e6)
}

View file

@ -78,15 +78,18 @@ func TestReadBulkRequest_Success(t *testing.T) {
{"create":{"_index":"filebeat-8.8.0"}}
{"@timestamp":"2023-06-06 04:48:12.735+01:00","message":"baz"}
{"index":{"_index":"filebeat-8.8.0"}}
{"message":"xyz","@timestamp":"2023-06-06T04:48:13.735Z","x":"y"}
{"message":"xyz","@timestamp":"1686026893735","x":"y"}
{"create":{"_index":"filebeat-8.8.0"}}
{"message":"qwe rty","@timestamp":"1686026893"}
`
timeField := "@timestamp"
msgField := "message"
rowsExpected := 3
timestampsExpected := []int64{1686026891735000000, 1686023292735000000, 1686026893735000000}
rowsExpected := 4
timestampsExpected := []int64{1686026891735000000, 1686023292735000000, 1686026893735000000, 1686026893000000000}
resultExpected := `{"@timestamp":"","log.offset":"71770","log.file.path":"/var/log/auth.log","_msg":"foobar"}
{"@timestamp":"","_msg":"baz"}
{"_msg":"xyz","@timestamp":"","x":"y"}`
{"_msg":"xyz","@timestamp":"","x":"y"}
{"_msg":"qwe rty","@timestamp":""}`
f(data, timeField, msgField, rowsExpected, timestampsExpected, resultExpected)
}

View file

@ -15,6 +15,8 @@ according to [these docs](https://docs.victoriametrics.com/victorialogs/quicksta
## tip
* FEATURE: [ElasticSearch bulk API](https://docs.victoriametrics.com/victorialogs/data-ingestion/#elasticsearch-bulk-api): accept timestamps in seconds in the ingested logs.
## [v0.31.0](https://github.com/VictoriaMetrics/VictoriaMetrics/releases/tag/v0.31.0-victorialogs)
Released at 2024-09-27

View file

@ -46,7 +46,15 @@ It is possible to push thousands of log lines in a single request to this API.
If the [timestamp field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#time-field) is set to `"0"`,
then the current timestamp at VictoriaLogs side is used per each ingested log line.
Otherwise the timestamp field must be in the [ISO8601](https://en.wikipedia.org/wiki/ISO_8601) or [RFC3339](https://www.rfc-editor.org/rfc/rfc3339) format.
Otherwise the timestamp field must be in one of the following formats:
- [ISO8601](https://en.wikipedia.org/wiki/ISO_8601) or [RFC3339](https://www.rfc-editor.org/rfc/rfc3339).
For example, `2023-06-20T15:32:10Z` or `2023-06-20 15:32:10.123456789+02:00`.
If timezone information is missing (for example, `2023-06-20 15:32:10`),
then the time is parsed in the local timezone of the host where VictoriaLogs runs.
- Unix timestamp in seconds or in milliseconds. For example, `1686026893` (seconds) or `1686026893735` (milliseconds).
For example, `2023-06-20T15:32:10Z` or `2023-06-20 15:32:10.123456789+02:00`.
If timezone information is missing (for example, `2023-06-20 15:32:10`), then the time is parsed in the local timezone of the host where VictoriaLogs runs.