Commit graph

327 commits

Author SHA1 Message Date
Aliaksandr Valialkin
bb161497cf app/vmselect/promql: improve the accuracy of buckets_limit(k, buckets) function
Now it properly merges the bucket with the previous bucket after deletion.
2020-07-24 17:07:49 +03:00
Aliaksandr Valialkin
994fa2f3bf app/vmselect/promql: add buckets_limit(k, buckets) function, which limits the number of buckets per time series to k
This function works with both Prometheus-style and VictoriaMetrics-style buckets.
The function removes buckets with the lowest values in order to reserve the highest precision.
The function is useful for building heatmaps in Grafana from too big number of buckets.
2020-07-24 16:13:53 +03:00
Aliaksandr Valialkin
e151c5c644 app/vmselect: fix tests for rate_over_sum 2020-07-24 02:35:28 +03:00
Aliaksandr Valialkin
3107c633e3 app/vmselect/promql: typo fix after 3e557c9861 2020-07-24 02:15:58 +03:00
Aliaksandr Valialkin
3e557c9861 app/vmselect/promql: add rate_over_sum(m[d]) function to MetricsQL, which returns rate over sum of m values over d duration
Something like `sum_over_time(m[d]) / d`, but more accurate.
2020-07-24 01:17:42 +03:00
Aliaksandr Valialkin
b1f6843bd0 app/vmselect/promql: allow setting [d] window smaller than the interval between raw points for avg_over_time
This makes `avg_over_time` behavior consistent with `sum_over_time` and `count_over_time` behaviors.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/636
2020-07-23 22:25:43 +03:00
Aliaksandr Valialkin
039c9d2441 lib/storage: respect -search.maxQueryDuration when searching for time series in inverted index
Previously the time spent on inverted index search could exceed the configured `-search.maxQueryDuration`.
This commit stops searching in inverted index on query timeout.
2020-07-23 21:21:42 +03:00
Aliaksandr Valialkin
2a45871823 lib/storage: add more fine-grained pace limiting for search 2020-07-23 19:26:08 +03:00
Aliaksandr Valialkin
461481fbdf app/vmselect/netstorage: protect from too smart compiler, which may break memory usage optimization in ProcessSearchQuery 2020-07-23 17:54:01 +03:00
Aliaksandr Valialkin
e79de9774b app/vmselect: typo fix after 34563916f7 2020-07-23 14:12:28 +03:00
Aliaksandr Valialkin
34563916f7 app/vmselect: reduce memory usage when querying big number of time series with long labels
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/646
2020-07-23 13:53:52 +03:00
Aliaksandr Valialkin
4f526cc816 app/vmselect/prometheus: support d, w and y suffixes for durations passed to step in /api/v1/query_range like Prometheus does
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/641
2020-07-22 16:26:18 +03:00
Aliaksandr Valialkin
dfb113f175 app/vmselect/netstorage: reduce memory allocations when unpacking time series data by using a pool for unpackWork entries
This should slightly reduce load on GC when processing queries that touch big number of time series.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/646 according to the provided memory profile there.
2020-07-22 15:03:57 +03:00
Aliaksandr Valialkin
caa2952aa6 app/vmselect: take into account the time spent in wait queue before query execution as time spent on the query 2020-07-21 19:00:09 +03:00
Aliaksandr Valialkin
e00cfc854d app/vmselect/promql: skip the first value in time series passed to increase() if it exceeds by more than 10x the delta between the next value and the first value
This should prvent from inflated `increase()` results for time series that start from big initial values.
Such cases may occur when a label value changes in a metric without counter reset.
2020-07-21 17:24:10 +03:00
Aliaksandr Valialkin
b9c8f6bf34 app/vmselect: log the total available memory for concurrent requests on not enough memory errors
This should simplify root cause analysis
2020-07-20 19:51:40 +03:00
Aliaksandr Valialkin
5a4675c528 app/vmselect/prometheus: do not return time series with empty list of datapoints from /api/v1/query_range
This matches Prometheus behaviour.

This should fix https://github.com/jacksontj/promxy/issues/329
2020-07-20 15:31:21 +03:00
Aliaksandr Valialkin
ecb1b2564a app/vmselect/promql: add mode() aggregate function 2020-07-20 15:31:20 +03:00
Aliaksandr Valialkin
b35cb293f5 lib/httpserver: log remote address in error message from httpserver.Errorf
This should improve detection of the root cause of errors.
Thanks to Anant for the idea.
2020-07-20 14:11:22 +03:00
Aliaksandr Valialkin
1c641037e8 app/vmselect/promql: add mode_over_time(m[d]) function
See https://en.wikipedia.org/wiki/Mode_(statistics) and https://stackoverflow.com/questions/61134078/promql-query-to-return-the-value-from-a-range-vector-which-occurs-maximum-no-of
2020-07-17 18:28:45 +03:00
Aliaksandr Valialkin
6b5ad535ae app/vmselect/promql: optimize group(rollup(m)) calculations 2020-07-17 16:47:16 +03:00
Aliaksandr Valialkin
8949d65ad1 app/vmselect/promql: check that any() doesn't touch metric name 2020-07-17 16:23:21 +03:00
Aliaksandr Valialkin
aa5d88055d app/vmselect/promql: add group() aggregate function to MetricsQL
This function has been added in Prometheus 2.20. See https://github.com/prometheus/prometheus/pull/7480
2020-07-17 15:17:55 +03:00
Aliaksandr Valialkin
df01836818 app/vmselect/promql: keep all labels for time series from any() call 2020-07-17 15:17:54 +03:00
Aliaksandr Valialkin
e4e1cd1de2 app/vmselect: fix nil pointer dereference panic when unsuccessfully querying vmstorage 2020-07-16 19:15:43 +03:00
Aliaksandr Valialkin
25f80d320b app/vmselect/prometheus: do not adjust last points in time series with timestamps exceeding the current time
Such timestamps usually mean that the query contains `offset`.
See https://github.com/VictoriaMetrics/VictoriaMetrics/issues/625
2020-07-14 12:52:16 +03:00
Aliaksandr Valialkin
55d83e777d app/vmselect/prometheus: minimize the diff for the change 1033dc7e2a over 619b0a25c9 2020-07-13 21:40:38 +03:00
faceair
1033dc7e2a
fix empty response template (#617) 2020-07-13 21:31:19 +03:00
Aliaksandr Valialkin
9c350bc20d app/vmselect/promql: add missing tests for ifnot binary operation 2020-07-09 13:24:06 +03:00
Aliaksandr Valialkin
256fd9a87e app/vmselect/promql: refactor implementations for and and unless binary operations, so they are closer to or implementation 2020-07-09 13:05:55 +03:00
Aliaksandr Valialkin
2d9b3ad5b3 app/vmselect/promql/active_queries.go: simplify code a bit by inlining getNextActiveQueryID function 2020-07-09 11:18:30 +03:00
Aliaksandr Valialkin
742da690f4 app/vmselect: add /api/v1/status/active_queries page with the list of currently running queries
This is a follow-up for https://github.com/VictoriaMetrics/VictoriaMetrics/pull/598

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/pull/575
2020-07-08 18:55:38 +03:00
DexterZhang
99f54e44ff
feat(vmselect): add current running query list, add ability for getting the running query info and killing running query for master branch (#598) 2020-07-08 18:52:55 +03:00
Aliaksandr Valialkin
acf828a759 app/vmselect/prometheus: small fixes on top of 8bb762124a 2020-07-05 18:17:06 +03:00
faceair
8bb762124a
fix adjust last points avoid influence earlier value (#606) 2020-07-05 17:56:54 +03:00
Aliaksandr Valialkin
5d3db3ff7c app/vmselect: add interpolate function for filling gaps with linearly interpolated values
See https://stackoverflow.com/q/62565021/274937 for details
2020-07-02 14:54:21 +03:00
Aliaksandr Valialkin
84a37098ed app/vmstorage: add -denyQueriesOutsideRetention command-line flag for denying queries outside the configured retention
VictoriaMetrics returns `503 Service Unavailable` http error for requests with time ranges outside the configured retention
if `-denyQueriesOutsideRetention` command-line flag is set.
2020-07-01 00:21:44 +03:00
Aliaksandr Valialkin
7c2c8b2981 all: use errors.As for inspecting errors that implement httpserver.ErrorWithStatusCode 2020-07-01 00:04:34 +03:00
Aliaksandr Valialkin
d5dddb0953 all: use %w instead of %s for wrapping errors in fmt.Errorf
This will simplify examining the returned errors such as httpserver.ErrorWithStatusCode .
See https://blog.golang.org/go1.13-errors for details.
2020-06-30 23:05:11 +03:00
Aliaksandr Valialkin
05d754d7bb app/vmselect/netstorage: reset big result values every 10 seconds instead of after processing every time series
This should reduce GC pressure when processing time series with big number of rows
2020-06-24 19:38:39 +03:00
nicbaz
774f7ca1c1
vmselect: fix label_replace when mismatch (#579)
As per documentation on `label_replace` function: "If the regular
expression doesn't match then the timeseries is returned unchanged".

Currently this behavior is not enforced, if a regexp on an existing
tag doesn't match then the tag value is copied as-is in the destination
tag. This fix first checks that the regular expression matches the
source tag before applying anything.

Given the current implementation, this fix also changes the behavior
of the **MetricsQL** `label_transform` function which does not
document this behavior at the moment.
2020-06-23 23:50:33 +03:00
Aliaksandr Valialkin
7209d58fbd app/vmselect/netstorage: increase concurrency when processing small number of time series with big number of data points per each time series
Previously VictoriaMetrics was processing up to 32 time series in a single goroutine.
This could be slow if each time series contains big number of data points (10M+ or more), since only a single CPU core could be loaded with work,
while other CPU cores were idle. Fix this by launching GOMAXPROCS workers for time series processing.

This should help with https://github.com/VictoriaMetrics/VictoriaMetrics/issues/572
2020-06-23 22:46:15 +03:00
Aliaksandr Valialkin
7532dbcdf5 app/vmselect/promql: properly override label values from group_left and group_right lists like Prometheus does
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/577
2020-06-21 16:33:01 +03:00
Aliaksandr Valialkin
88e1b7d144 app/vmselect/promql: fill gaps on right side with values from left side of or operator in the same way as Prometheus does
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/552
2020-06-18 23:05:16 +03:00
Aliaksandr Valialkin
dbd8beccfa app/vmselect/promql: add ascent_over_time(m[d]) and descent_over_time(m[d]) functions
These functions could be useful in GPS tracking apps for calculating the summary for height gain/loss
over the given duration `d`.
2020-05-21 12:07:48 +03:00
Aliaksandr Valialkin
70ea4e28a7 app/vmselect/promql: update numbers after the upgrade of github.com/VictoriaMetrics/metrics from v1.11.2 to v1.11.3 2020-05-20 03:06:23 +03:00
Aliaksandr Valialkin
538fdfe133 app/vmselect/promql: move common code from aggrFuncOutliersK and newAggrFuncRangeTopK into getRangeTopKTimeseries 2020-05-19 16:11:14 +03:00
Aliaksandr Valialkin
f52769f6ee app/vmselect/promql: fix outilersk calculations 2020-05-19 14:44:53 +03:00
Aliaksandr Valialkin
a441cdd1d9 app/vmselect/promql: add outliersk(N, m) aggregate function for anomaly detection across groups of similar time series 2020-05-19 13:52:36 +03:00
Aliaksandr Valialkin
0afd48d2ee lib: extract common code for returning fast unix timestamp into lib/fasttime 2020-05-14 23:02:07 +03:00
Aliaksandr Valialkin
7882a0dbbf app/vmselect/promql: suppress "SA4006: this value of dstValues is never used" error in golangci-lint 2020-05-13 11:47:08 +03:00
Aliaksandr Valialkin
faf92a0965 app/vmselect/promql: fix any(..) calculations - return all the data points instead of the first one 2020-05-12 20:36:42 +03:00
Aliaksandr Valialkin
cc311e20fe app/vmselect/promql: add any(x) by (y) aggregate function, which returns any time series from q for each group y 2020-05-12 19:45:56 +03:00
Aliaksandr Valialkin
574289c3fb app/vmselect/promql: support for sum(x) by (y) limit N syntax in order to limit the number of output time series after aggregation 2020-05-12 19:45:54 +03:00
Aliaksandr Valialkin
57407cca83 app/vmselect/promql: remove -search.maxPointsPerTimeseries command-line flag
Limit the estimated time series count after aggregation with grouping by the number of source time series.
2020-04-29 00:20:04 +03:00
Aliaksandr Valialkin
4e4f57b121 lib/metricsql: move it to a separate repository - github.com/VictoriaMetrics/metrics 2020-04-28 15:28:22 +03:00
Aliaksandr Valialkin
17d96e4503 app/vmselect: add -search.estimatedSeriesCountAfterAggregation command-line flag for tuning the probability of OOMs vs false-positive not enough memory errors 2020-04-28 12:52:37 +03:00
Aliaksandr Valialkin
b4afe562c1 lib/storage: postpone reading data from blocks during search
This eliminates the need for storing block data into temporary files on a single-node VictoriaMetrics
during heavy queries, which touch big number of time series over long time ranges.

This improves single-node VM performance on heavy queries by up to 2x.
2020-04-27 11:45:24 +03:00
Aliaksandr Valialkin
fcf57f9883 app/vmselect/netstorage: substitute sorting packedTimeseries with the natural order of the fetched blocks
This should minimize the number of disk seeks when reading data from temporary file.
2020-04-26 16:26:23 +03:00
Aliaksandr Valialkin
a596aec82c app/vmselect: fix description for -search.resetCacheAuthKey 2020-04-24 19:45:50 +03:00
Aliaksandr Valialkin
364db13c9c app/vmselect: add /api/v1/status/tsdb page with useful stats for locating root cause for high cardinality issues
See https://prometheus.io/docs/prometheus/latest/querying/api/#tsdb-stats

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/425
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/268
2020-04-22 22:03:43 +03:00
Aliaksandr Valialkin
9ebc937685 app/vmselect: add -search.minStalenessInterval command-line flag for removing gaps on graphs built from time series with irregular duration between samples
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/426
2020-04-20 19:42:15 +03:00
Aliaksandr Valialkin
fe57d46687 app/vmselect: merge -search.maxLookback and -search.maxStalenessInterval flags, since it has been appeared they have identical purpose :(
Leave both flags for backwards compatibility reasons.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/209
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/426
2020-04-20 19:26:31 +03:00
Aliaksandr Valialkin
f41e6a7bd9 app/vmselect: properly apply -search.maxLookback to queries sent to /api/v1/query 2020-04-17 12:30:11 +03:00
Aliaksandr Valialkin
071fdf5518 lib/logger: add WARN level for logging expected errors such as invalid user queries 2020-04-15 20:50:26 +03:00
Aliaksandr Valialkin
6f7f64f757 app/vmselect: handle timestamp(metric offset X) the same way as Prometheus does
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/415
2020-04-15 12:01:00 +03:00
Aliaksandr Valialkin
f9b24d4899 app/vmselect/promql: keep metric name after applying first_over_time and last_over_time functions 2020-04-04 14:54:13 +03:00
Aliaksandr Valialkin
318326c309 lib/storage: properly handle {label=~"foo|"} filters as Prometheus does
Such filters must match all the time series with `label="foo"` plus all the time series without `label`

Previously only time series with `label="foo"` were matched.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/395
2020-03-31 00:48:18 +03:00
Aliaksandr Valialkin
0c485f14d1 app/vmselect/prometheus: allow passing relative time to start, end and time args of /api/v1/* queries 2020-03-29 21:57:14 +03:00
Aliaksandr Valialkin
2ebf7d86ff app/vmselect/prometheus: code simplification: (d.Seconds()/1e3) -> d.Milliseconds() 2020-03-29 21:50:28 +03:00
Aliaksandr Valialkin
3c7c71a49c app/vmselect: adjust label_map() handling for corner cases
The following corner cases now supported:
* label_map(q, "label", "", "foo") - adds `label="foo"` to series with missing `label`
* label_map(q, "label", "foo", "") - removes `label="foo"` from series

All the unmatched labels are kept unchanged.
2020-03-13 18:45:03 +02:00
Aliaksandr Valialkin
fdc2a9d1d7 app/vmselect: add label_map(q, label, srcValue1, dstValue1, ... srcValueN, dstValueN) function to MetricsQL
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/369
2020-03-12 19:13:47 +02:00
Aliaksandr Valialkin
8023ad7dbd app/vmselect: add -search.maxStalenessInterval for tuning Prometheus data model closer to Influx-style data model 2020-03-11 16:43:34 +02:00
Aliaksandr Valialkin
e17702fada app/vmselect: add optional max_rows_per_line query arg to /api/v1/export
This arg allows limiting the number of data points that may be exported on a single line.
2020-03-10 21:45:56 +02:00
Aliaksandr Valialkin
c5f894b361 Makefile: add build and test rules with enabled race detector. These rules have -race suffix
Fix also `unsafe pointer conversion` errors detected by Go1.14. See https://golang.org/doc/go1.14#compiler .
2020-03-05 12:03:38 +02:00
Aliaksandr Valialkin
7d178a40bd app/vmselect/prometheus: do not add __name__!= filter when searching for all the matching metric names via /api/v1/label/__name__/values with non-empty label filter
This should reduce query time.
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/343
2020-02-28 23:35:55 +02:00
Aliaksandr Valialkin
5701b2f7bb app/vmselect/prometheus: properly pass filter for labelName=__name__ in labelValuesWithMatches
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/343
2020-02-28 12:18:14 +02:00
Aliaksandr Valialkin
18af31a4c2 all: properly split vm_deduplicated_samples_total among cluster components
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/345
2020-02-27 23:48:07 +02:00
Aliaksandr Valialkin
ae51300973 app/vmselect/promql: properly take into account the first datapoint when calculating rollup_candlestick
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/309
2020-02-24 13:24:30 +02:00
Aliaksandr Valialkin
e65ec88779 app/vmselect/promql: do not take into account values outside the current window in rollup_candlestick
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/309
2020-02-23 18:03:57 +02:00
Yaroslav
a6d0645539
fix rollupOpen(), rollupHigh(), rollupLow() functions (#328) 2020-02-23 18:01:53 +02:00
Aliaksandr Valialkin
2f55cabaa4 app/vmselect/promql: log when rollupResult cache is cleared 2020-02-21 20:07:01 +02:00
Aliaksandr Valialkin
ec9bf39b5b app/vmselect: add -search.cacheTimestampOffset command-line flag
This flag can be used for removing gaps on graphs if the difference between the current time
and the timestamps from the ingested data exceeds 5 minutes.

This is the case when the time between data sources and VictoriaMetrics is improperly synchronized.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/312
2020-02-21 13:58:06 +02:00
Aliaksandr Valialkin
539139391c app/vmselect: add /internl/resetRollupResultCache handler for resetting response cache
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/312
2020-02-21 13:58:05 +02:00
Aliaksandr Valialkin
7b1c7051a3 app/vmselect: add sort_by_label(q, label) and sort_by_label_desc(q, label) functions
This is implementation of https://github.com/prometheus/prometheus/pull/1533 for VictoriaMetrics.
2020-02-13 17:01:37 +02:00
Aliaksandr Valialkin
a6c6a2debc app/vmselect/promql: do not add step to range end, since this hack became obsolete since commit 9e1119dab8 2020-02-05 21:22:19 +02:00
Aliaksandr Valialkin
78b62dee87 app/vmselect/promql: properly adjust time range for data to select
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/309
2020-02-05 21:22:18 +02:00
Aliaksandr Valialkin
366693b9f1 app/vmselect: unconditionally offset -step to rollup_candlestick. This makes results more consistent 2020-02-04 23:32:12 +02:00
Aliaksandr Valialkin
525101339e app/vmselect/promql: automatically apply offset -step to rollup_candlestick function in order to obtain the expected OHLC results
See https://github.com/VictoriaMetrics/VictoriaMetrics/issues/309
2020-02-04 23:24:35 +02:00
Aliaksandr Valialkin
ada6a3da8d app/vmselect/promql: adjust rollup_candlestick calculations to the exepcted results
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/309
2020-02-04 22:42:13 +02:00
Aliaksandr Valialkin
36a1ac8360 app/vmselect: take into account the time the requests wait in the queue if -search.maxConcurrentRequests is exceeded
This will prevent from excess CPU usage for timed out queries.
2020-02-04 16:15:08 +02:00
Aliaksandr Valialkin
834051e5b2 app/vmselect: add a placeholder for /api/v1/metadata, which could be requested by Grafana
See https://prometheus.io/docs/prometheus/latest/querying/api/#querying-metric-metadata

VictoriaMetrics doesn't collect any metadata for metrics, so just return empty response.
2020-02-04 15:53:47 +02:00
Aliaksandr Valialkin
42864bb52f all: do not clash flag description with back-quoted flag types
See https://golang.org/pkg/flag/#PrintDefaults for more details.
2020-02-04 15:46:52 +02:00
Aliaksandr Valialkin
90cf356ea1 app/vmselect/promql: adjust and and unless binary operator handling to be consistent with Prometheus 2020-01-31 18:52:38 +02:00
Aliaksandr Valialkin
c3d86eef96 all: add -dedup.minScrapeInterval command-line flag for data de-duplication
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/86
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/278
2020-01-31 01:16:57 +02:00
Aliaksandr Valialkin
d70ba7eb37 lib/fs: optimize small reads for ReaderAt.MustReadAt by reading from memory-mapped space instead of reading from file descriptor
This should improve performance when reading many small blocks.
2020-01-30 15:09:05 +02:00
Aliaksandr Valialkin
a462355b2f app/vmselect/promql: add keep_next_value(q) for filling gaps with the next non-empty value 2020-01-29 00:48:04 +02:00
Aliaksandr Valialkin
adbbc4fa1a app/vmselect/promql: return expected results from increase() over the beginning of time series, which start from big value
Examples for such counters: OS-level counters for network or cpu stats.
2020-01-28 16:30:11 +02:00
Aliaksandr Valialkin
2392a359e1 app/vmselect/promql: fix panic on a single zero vmrange bucket in prometheus_buckets() function
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/296
2020-01-27 18:04:55 +02:00
Aliaksandr Valialkin
680080887d all: consistently log durations in seconds with millisecond precision
This should improve logs readability
2020-01-22 18:28:27 +02:00