Commit graph

1260 commits

Author SHA1 Message Date
Aliaksandr Valialkin
a8bcc3c276
lib/promscrape: reduce memory usage if -promscrape.noStaleMarkers command-line flag is passed
Do not store in memory the response from the last scrape per each target if -promscrape.noStaleMarkers option is enabled.
This should reduce memory usage when the scraped targets return large responses.
2021-10-22 13:22:08 +03:00
Roman Khavronenko
e0f21d6000
vmalert: correctly calculate alert ID including extra labels (#1734)
Previously, ID for alert entity was generated without alertname or groupname.
This led to collision, when multiple alerting rules within the same group
producing same labelsets. E.g. expr: `sum(metric1) by (job) > 0` and
expr: `sum(metric2) by (job) > 0` could result into same labelset `job: "job"`.

The issue affects only UI and Web API parts of vmalert, because alert ID is used
only for displaying and finding active alerts. It does not affect state restore
procedure, since this label was added right before pushing to remote storage.

The change now adds all extra labels right after receiving response from the datasource.
And removes adding extra labels before pushing to remote storage.

Additionally, change introduces a new flag `Restored` which will be displayed in UI
for alerts which have been restored from remote storage on restart.
2021-10-22 12:31:35 +03:00
Yury Molodov
3ccc37a98b
vmui: query history (#1732)
* feat: add query history

* fix: change detect keyUp for nav query history

* feat: set default query history

* app/vmselect/vmui: `make vmui-update`

Co-authored-by: Aliaksandr Valialkin <valyala@victoriametrics.com>
2021-10-22 12:29:27 +03:00
Aliaksandr Valialkin
5705f4b6d1
lib/httpserver: expose command-line flags at /flags page
This should simplify debugging.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1695
2021-10-20 00:46:54 +03:00
Roman Khavronenko
ca49853664
vmalert: make group.ID() thread-safe (#1726)
Commit fixes potential race condition when group update
and generating of ID() happens simultaneously.

Signed-off-by: hagen1778 <roman@victoriametrics.com>
2021-10-19 16:45:52 +03:00
Roman Khavronenko
627224d493
vmalert: properly init SIGHUP listener before starting group manager (#1725)
Regression was introduced during code refactoring. It potentially
could lead to situation when SIGHUP signals were ignored while
vmalert was still busy with initing group manager.

Signed-off-by: hagen1778 <roman@victoriametrics.com>
2021-10-19 16:36:08 +03:00
Aliaksandr Valialkin
0a1982b294
app/vmauth: fix metric name prefixes: vmagent -> vmauth 2021-10-19 15:30:16 +03:00
Yury Molodov
744f6a98eb
vmui: features (#1711)
* feat: initial uPlot graph

* feat: add zoom/pan for graph

* fix: add zoom by ctrl/mac

* fix: remove unused code

* feat: add toggle cache for fetch

* feat: add fix y-axis limits

* fix: stop point events while panning

* fix: change getting cursor position when scaling

* feat: add cursor tooltip to graph

* fix: uninstall chart.js

* fix: change link for create an issue

* fix: set default cache value to true

* app/vmalert: follow-up after 0e2486df56

* docs/CHANGELOG.md: document 5416e18007

* app/vmui: `make vmui-update`

Co-authored-by: Aliaksandr Valialkin <valyala@victoriametrics.com>
2021-10-18 15:19:01 +03:00
Roman Khavronenko
44a88b7458
vmalert: remove extra / from path in WEB interface (#1717)
The extra `/` may cause issues when additional path prefixes
are configured. Also, removing it makes it consistent
with the rest of declarations.

Signed-off-by: hagen1778 <roman@victoriametrics.com>
2021-10-18 15:19:01 +03:00
Roman Khavronenko
93d4ef1239
vmctl: follow-up after 95d1d38595 (#1718)
Signed-off-by: hagen1778 <roman@victoriametrics.com>
2021-10-18 15:19:01 +03:00
Aliaksandr Valialkin
9273b83cd0
app/vmalert: follow-up after 0e2486df56 2021-10-18 15:08:42 +03:00
Miro Prasil
c0853d4bf8
vmctl influx convert bool to number (#1714)
vmctl: properly convert influx bools into integer representation

When using vmctl influx, the import would fail importing boolean fields
with:

```
failed to convert value "some".0 to float64: unexpected value type true
```

This converts `true` to `1` and `false` to `0`.

Fixes #1709
2021-10-18 14:59:17 +03:00
Alexander Rickardsson
0e1dbcd039
vmalert: add disablePathAppend to remote read (#1712)
* vmalert: add disablePathAppend to remoteRead

* docs: add docs for remoteRead.disablePathAppend
2021-10-18 14:59:17 +03:00
Alexander Rickardsson
63571e1334
vmalert: Redact passwords from error messages (#1713) 2021-10-18 14:59:17 +03:00
Aliaksandr Valialkin
ed994873fd
app/vmselect/promql: randomize the static selection of time series returned from limitk()
Sort series by a hash calculated from the series labels. This should guarantee "random" selection of the returned time series.
Previously the selection could be biased, since time series were sorted alphabetically by label names and label values.
2021-10-16 21:16:34 +03:00
Aliaksandr Valialkin
da0adf446d
app/vmselect/promql: typo fix in comment: didsn't -> didn't 2021-10-16 13:00:10 +03:00
Aliaksandr Valialkin
0452a8d4e8
lib/promscrape: store the full response in stream parsing mode in scrapeWork.lastScrape byte slice
This allows sending staleness marks and properly calculate scrape_series_added metric in stream parsing mode
at the cost of the increased memory usage, since now the potentially big response is kept
in the lastScrape byte slice per each scrapeWork.

In practice the memory usage increase shouldn't be big, since the response size
is usually much smaller than the parsed metrics from this response after the relabeling,
which usually adds a big pile of target-specific labels per each metric.
2021-10-15 15:26:24 +03:00
Aliaksandr Valialkin
25421fa2ae
lib/promscrape: add -promscrape.minResponseSizeForStreamParse command-line option for automatic switching to stream parsing mode when scraping targets with big responses
This should reduce memory usage when vmagent scrapes targets with non-uniform response sizes.
This is common case in Kubernetes monitoring.
2021-10-14 12:30:55 +03:00
Aliaksandr Valialkin
bee130cc78
lib/promscrape: return error if sample_limit or series_limit options are set when stream parsing mode is enabled 2021-10-14 12:30:54 +03:00
Roman Khavronenko
f393145843
Adjust http.Transport.MaxIdleConns setting for vmauth/vmalert services (#1704)
* vmalert: adjust `http.Transport.MaxIdleConns` value accordingly to `http.Transport.MaxIdleConnsPerHost`

`http.Transport.MaxIdleConnsPerHost` setting is controlled by `datasource.maxIdleConnections` flag,
while `http.Transport.MaxIdleConns` is inherited from DefaultTransport and is equal to `100`.
The fix adjusts `http.Transport.MaxIdleConns` value if it is lower than `http.Transport.MaxIdleConnsPerHost`.

Signed-off-by: hagen1778 <roman@victoriametrics.com>

* vmauth: adjust `http.Transport.MaxIdleConns` value accordingly to `http.Transport.MaxIdleConnsPerHost`

`http.Transport.MaxIdleConnsPerHost` setting is controlled by `maxIdleConnsPerBackend` flag,
while `http.Transport.MaxIdleConns` is inherited from DefaultTransport and is equal to `100`.
The fix adjusts `http.Transport.MaxIdleConns` value if it is lower than `http.Transport.MaxIdleConnsPerHost`.

Signed-off-by: hagen1778 <roman@victoriametrics.com>
2021-10-13 19:23:52 +03:00
Roman Khavronenko
14995eece6
vmalert: add Source link to alerts UI (#1701)
The source link is controlled by `external.url` and `external.alert.source`
flags, in the same way as for alertmanager notifications.
The source link is added to Alerts list view, and specific Alert view.
2021-10-13 15:26:57 +03:00
Aliaksandr Valialkin
aeedfe2fe2
app/vmagent: expose -promscrape.config contents at /config page as Prometheus does
See https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1695
2021-10-12 16:27:37 +03:00
Aliaksandr Valialkin
48c3668b3d
app/vmselect/promql: add atan2 binary operator, which is going to be added in Prometheus 2.31
See https://github.com/prometheus/prometheus/pull/9248
2021-10-11 21:17:23 +03:00
Aliaksandr Valialkin
b3d28f3872
app/vmselect/promql: add missing trigonometric functions, which are going to be added in Prometheus 2.31
See https://github.com/prometheus/prometheus/issues/9233
2021-10-11 21:17:21 +03:00
Aliaksandr Valialkin
f1e1d20ac4
app/vmselect/promql: consistently return the same set of time series from limitk() function
This is the expected behaviour by most users.
2021-10-08 19:55:29 +03:00
Aliaksandr Valialkin
cf87050d83
app/vmui: follow-up after 7bfb44113e
* Run `vmui-update`
* Document the changes in README.md and CHANGELOG.md
2021-10-08 15:11:09 +03:00
Yury Molodov
41a83a1cc6
vmui: use uPlot as default engine for graph (#1683)
* feat: initial uPlot graph

* feat: add zoom/pan for graph

* fix: add zoom by ctrl/mac

* fix: remove unused code
2021-10-08 15:11:07 +03:00
Aliaksandr Valialkin
4fddcf4c83
app/{vminsert,vmstorage}: follow-up after a171916ef5
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/269
2021-10-08 14:09:51 +03:00
Nikolay
a171916ef5
Adds read-only mode for vmstorage node (#1680)
* adds read-only mode for vmstorage
https://github.com/VictoriaMetrics/VictoriaMetrics/issues/269

* changes order a bit

* moves isFreeDiskLimitReached var to storage struct
renames functions to be consistent
change protoparser api - with optional storage limit check for given openned storage

* renames freeSpaceLimit to ReadOnly
2021-10-08 12:52:56 +03:00
Aliaksandr Valialkin
cbf01f7384
app/vmselect/promql: substitute rollupFuncsCannotAdjustWindow with rollupFuncsCanAdjustWindow
The list of functions, which can adjust lookbehind window is more limited than the rest of functions,
so it is better from maintainability and readability PoV using the allowlist instead of blocklist.
2021-10-07 13:19:22 +03:00
Aliaksandr Valialkin
7e6ec0bbe9
app/vmselect/promql: return back the behaviour for deriv() function when the lookbehind window doesnt contain enough points
It is expected that the `deriv(m[d])` returns non-empty value if the lookbehind window `d`
contains less than 2 samples in the same way as `rate()` does.

This is a follow-up after 3e084be06b .
2021-10-07 12:55:58 +03:00
Roman Khavronenko
0a8804d6aa
app/vmselect: make predict_linear and deriv compatible with Prometheus (#1681)
Previously, `predict_linear` returned slightly different results comparing
to Prometheus. The change makes linear regression algorithm compatible
with Prometheus.

`deriv` was excluded from the list of functions which can adjust the time
window for the same reasons.
2021-10-07 12:55:56 +03:00
Aliaksandr Valialkin
64b6f3f1c8
app/vminsert: fix uneven distribution of time series among storage nodes
Use distinct seed for distribution hash calculations on the second level of vminsert nodes.

See https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1672
2021-10-07 12:22:39 +03:00
Aliaksandr Valialkin
89700433cc
docs/vmagent.md: update docs after 3e9a939a990c8b608414388c96f68eb062364ae7 2021-10-05 10:23:37 +03:00
Aliaksandr Valialkin
a9420317d9
docs/vmagent.md: document how to write data to Kafka 2021-09-30 17:46:23 +03:00
Aliaksandr Valialkin
ec6eb03d65
app/vmselect/promql: add topk_last and bottomk_last functions 2021-09-30 13:23:27 +03:00
Roman Khavronenko
95ddfda894
app/vmselect: fix binary comparison func (#1667)
The fix makes the binary comparison func to check for NaNs
before executing the actual comparison. This prevents VM
to return values for non-existing samples for expressions
which contain bool comparisons. Please see added test
for example.
2021-09-30 13:04:10 +03:00
Roman Khavronenko
57eb5ef194
app/vmselect: fix testRowsEqual func NaN checks (#1666)
It appeared, that `testRowsEqual` NaN comparison was incorrect.
The fix caused some tests to fail. Please see the change and
tests updated.

Signed-off-by: hagen1778 <roman@victoriametrics.com>
2021-09-30 13:04:08 +03:00
Aliaksandr Valialkin
d51118e226
app/vminsert: document that -relabelConfig is reloaded on SIGHUP signal 2021-09-29 21:17:25 +03:00
Aliaksandr Valialkin
8dcf814c48
app/{vmbackup,vmrestore}: switch from gcs://... to gs://... urls for backups to GCS
The `gs://` urls are commonly used, so prefer them instead of `gcs://` urls,
while leaving support for `gcs://` urls for backwards compatibility.
2021-09-29 12:12:37 +03:00
Aliaksandr Valialkin
f35f9188bc
docs/vmagent.md: update Telegraf config in the section about Kafka 2021-09-29 12:01:46 +03:00
Aliaksandr Valialkin
16d7ee7b44
docs/vmagent.md: add docs about reading metrics from Kafka 2021-09-29 01:46:40 +03:00
Yury Molodov
a6037a8ce8
vmui: fixed bug with time range (time zone) (#1661)
* fix: set date in query string in utc format

* app/vmselect: `make vmui-update`

Co-authored-by: Aliaksandr Valialkin <valyala@victoriametrics.com>
2021-09-29 01:01:27 +03:00
Nikolay
6ca31acb76
changes vmagent api (#1656)
* changes vmagent api
adds auth.Token to promremotewrite InsertHandlerReader
changes remoteWrite client constructor, allows to use multiple remoteWriteUrl schemes, like kafka://
changes url path concatenation for tenant remoteWrite

Update app/vmagent/remotewrite/client.go

Co-authored-by: Aliaksandr Valialkin <valyala@victoriametrics.com>

* Update app/vmagent/remotewrite/remotewrite.go

* Apply suggestions from code review

Co-authored-by: Aliaksandr Valialkin <valyala@victoriametrics.com>
2021-09-29 00:55:14 +03:00
Roman Khavronenko
444a8882f4
app/vmselect: disable time-window adjustment for min/max_over_time funcs (#1658)
Adjustment results into discrepancy between Prometheus and VM on time windows
smaller than scrape interval.

Signed-off-by: hagen1778 <roman@victoriametrics.com>
2021-09-29 00:43:54 +03:00
Roman Khavronenko
f46f847531
app/vmselect: always return zero for stddev func if there is only one value (#1659)
The fix will always return zero if received set of items consists of one
element only, which also means no deviation.

Signed-off-by: hagen1778 <roman@victoriametrics.com>
2021-09-29 00:42:19 +03:00
Roman Khavronenko
72468c9d75
app/vmselect: return NaN instead of 0 for empty value sets (#1660)
The change affects `count/stddev/stdvar_over_time` funcs and makes
them to return NaN instead of zero when there is no datapoints
in a time window.
This is needed for improving compatibility with Prometheus.

Signed-off-by: hagen1778 <roman@victoriametrics.com>
2021-09-29 00:38:24 +03:00
Roman Khavronenko
c8ac8f44b2
app/vmselect: rm quantile_over_time fast-path optimisations (#1662)
The removed fast path optimisations weren't consistent with
`quantile` function behavior and results into discrepancy.
Specifically, results didn't match in cases when:
* 0 < phi > 1;
* values contain only one element.

Signed-off-by: hagen1778 <roman@victoriametrics.com>
2021-09-29 00:35:59 +03:00
Aliaksandr Valialkin
4e65bfcc00
app/{vminsert,vmagent}: add ability to ingest data via DataDog "submit metrics" API
See https://docs.datadoghq.com/api/latest/metrics/#submit-metrics

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/206
2021-09-29 00:12:26 +03:00
Yury Molodov
80df31b2ee
vmui: use Chart.js as default engine for graph (#1634)
* feat: add Plotly as default engine for graph

* fix: remove unused components

* feat: use Chart.js as default engine graph

* fix: correct styles for loader

* feat: add zoom/pan for chart

* feat: add height for chart

* fix: remove unused code

* fix: remove empty units from duration

* fix: change debounce for pan to 500ms

* fix: add utility for plugins register globally

* fix: optimize render graph

* feat: add buffer data for zoom

* fix: add limits for zoom in/out

* fix: change update data while zooming

* app/vmselect: `make vmui-update`

Co-authored-by: Aliaksandr Valialkin <valyala@victoriametrics.com>
2021-09-27 22:27:49 +03:00