Commit graph

341 commits

Author SHA1 Message Date
Aliaksandr Valialkin
414259f47b app/vmselect/promql: do not take into account buckets with negative counters in prometheus_buckets 2019-11-23 14:19:19 +02:00
Aliaksandr Valialkin
193d553f6d app/vmselect/promql: properly handle histogram_quantile(0, ...) with zero buckets 2019-11-23 14:02:25 +02:00
Aliaksandr Valialkin
4d76977745 app/vmselect/promql: transparently apply prometheus_buckets in histogram_quantile 2019-11-23 11:49:16 +02:00
Aliaksandr Valialkin
5f6f03c692 app/vmselect/promql: add prometheus_buckets function for converting the upcoming histogram buckets from github.com/VictoriaMetrics/metrics to Prometheus-compatible buckets 2019-11-23 00:21:56 +02:00
Aliaksandr Valialkin
77bb66a5be app/vmselect/promql: properly calculate integrate(q[d]) 2019-11-13 21:11:03 +02:00
Aliaksandr Valialkin
c33640664a app/vmselect/promql: use universal approach for determining maxByteSliceLen on 32-bit and 64-bit archs
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/235
2019-11-13 20:26:07 +02:00
Aliaksandr Valialkin
33abbec6b4 app/vmselect/promql: adjust memory limits calculations for incremental aggregate functions
Incremental aggregate functions don't keep all the selected time series in memory -
they keep only up to GOMAXPROCS time series for incremental aggregations.

Take into account that the number of time series in RAM can be higher if they are split
into many groups with `by (...)` or `without (...)` modifiers.

This should reduce the number of `not enough memory for processing ... data points` false
positive errors.
2019-11-08 19:37:43 +02:00
Aliaksandr Valialkin
4a8251feff app/vmselect/promql: add lag(q[d]) function, which returns the lag between the current timestamp and the timstamp for the last data point in q 2019-11-01 12:21:43 +02:00
Aliaksandr Valialkin
5b01b7fb01 all: add support for GOARCH=386 and fix all the issues related to 32-bit architectures such as GOARCH=arm
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/212
2019-10-17 18:27:49 +03:00
Aliaksandr Valialkin
99786c2864 app/vmselect/prometheus: add -search.maxLookback command-line flag for overriding dynamic calculations for max lookback interval
This flag is similar to `-search.lookback-delta` if set. The max lookback interval is determined dynamically
from interval between datapoints for each input time series if the flag isn't set.

The interval can be overriden on per-query basis by passing `max_lookback=<duration>` query arg to `/api/v1/query` and `/api/v1/query_range`.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/209
2019-10-15 21:37:17 +03:00
Aliaksandr Valialkin
a5302a6651 app/vmselect/promql: take into account the previous point when calculating max_over_time and min_over_time
This lines up with `first_over_time` function used in `rollup_candlestick`, so `rollup=low` always returns
the minimum value.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/204
2019-10-08 12:30:16 +03:00
Aliaksandr Valialkin
ba460f62e6 app/vmselect/promql: do not generate timestamps for NaN values in timestamp function according to Prometheus logic 2019-09-27 18:55:16 +03:00
Aliaksandr Valialkin
bd1cf053f6 app/vmselect/promql: add increases_over_time and decreases_over_time functions
`increases_over_time(q[d])` returns the number of `q` increases during the given duration `d`.
`decreases_over_time(q[d])` returns the number of `q` decreases during the given duration `d`.
2019-09-25 20:38:51 +03:00
Aliaksandr Valialkin
ee4585db33 app/vmselect/promql: properly handle subqueries like aggr_func(rollup_func(metric[window:step]))
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/184
2019-09-13 21:42:11 +03:00
Aliaksandr Valialkin
828e5f6d26 app/vmselect/promql: binary operation fixes according to Prometheus behaviour
The follosing issues were fixed:
- VictoriaMetrics could leave superflouos labels when using `on` or `ignoring` modifiers
- VictoriaMetrics could return `duplicate timeseries` error when using `group_left` or `group_right` with non-empty label list
2019-09-13 17:43:09 +03:00
Aliaksandr Valialkin
d0953e9f02 app/vmselect/promql: ignore grouping by destination label in count_values, since such a grouping is performed automatically 2019-09-04 19:59:02 +03:00
Aliaksandr Valialkin
f78ffe565f app/vmselect/promql: do not return artificial points beyond the last point in time series 2019-09-04 16:34:29 +03:00
Aliaksandr Valialkin
b08f085082 app/vmselect/promql: reset timeseries name on group_left and group_right as Prometheus does 2019-09-03 20:43:29 +03:00
Aliaksandr Valialkin
604a4312f9 all: port to FreeBSD on GOARCH=amd64 2019-08-28 01:46:09 +03:00
Aliaksandr Valialkin
e9db22a551 app/vmselect/promql: attempt to repair invalid bucket counts passed to histogram_quantile
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/136
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/154
2019-08-22 14:39:24 +03:00
Aliaksandr Valialkin
90a4b00b10 app/vmselect/promql: fix panic on -search.disableCache
Reset the cache if it is disabled instead of stopping, since it is stopped on graceful shutdown.
2019-08-21 17:12:01 +03:00
Aliaksandr Valialkin
491b1762c8 app/vmselect/promql: explain why empty timeseries arent removed in transformLabelValue 2019-08-21 11:29:41 +03:00
Aliaksandr Valialkin
99331606e1 app/vmselect/promql: pre-allocate memory for map for checking for duplicate timeseries
This should reduce memory allocations for big number of timeseries
2019-08-20 23:01:57 +03:00
Aliaksandr Valialkin
1101765adb app/vmselect/promql: add label_value(q, label_name) func, which returns numeric value labels with name label_name in q 2019-08-20 00:28:44 +03:00
Aliaksandr Valialkin
940349ccb9 app/vmselect/promql: independently track offset hints for tStart and tEnd
This should improve performance if timeseries starts or ends on the selected time range
2019-08-19 13:40:24 +03:00
Aliaksandr Valialkin
6ae4b4190f app/vmselect/promql: optimize search for timestamp boundaries in rollupConfig.Do
This should improve the performance of queries over big number of time series
with big number of output points.
2019-08-19 13:03:38 +03:00
Aliaksandr Valialkin
005aabd305 app/vmselect/promql: add scrape_interval(q[d]) function, which would return scrape interval for q over d 2019-08-18 21:08:15 +03:00
Aliaksandr Valialkin
218cb4623a app/vmselect/promql: hande comparisons with NaN similar to Prometheus
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/150
2019-08-18 00:25:58 +03:00
Aliaksandr Valialkin
dcce92c63c app/vmselect/promql: add lifetime(q[d]) function, which returns the lifetime of q over d in seconds.
This function is useful for determining time series lifetime.
`d` must exceed the expected lifetime of the time series, otherwise
the function would return values close to `d`.
2019-08-16 11:59:51 +03:00
Aliaksandr Valialkin
0cb66a8f95 app/vmselect/promql: fix corner-case calculation for ideriv 2019-08-16 11:59:50 +03:00
Aliaksandr Valialkin
1b5b9ced27 app/vmselect/promql: properly handle corner cases for rollup functions 2019-08-15 23:31:28 +03:00
Aliaksandr Valialkin
b8bbe92de1 app/vmselect/promql: store compressed results in the cache
This should increase rollup results cache capacity.
2019-08-14 02:32:16 +03:00
Aliaksandr Valialkin
8c2158af24 all: use workingsetcache instead of fastcache
This should reduce the amount of RAM required for processing time series
with non-zero churn rate.

The previous cache behavior can be restored with `-cache.oldBehavior` command-line flag.
2019-08-13 21:40:28 +03:00
Aliaksandr Valialkin
880b1d80b1 app/vmselect: optimize /api/v1/series by skipping storage data
Fetch and process only time series metainfo.
2019-08-04 23:00:46 +03:00
Aliaksandr Valialkin
000c154641 app/vmselect/promql: tune automatic window adjustement
Increase the windows adjustement for small scrape intervals,
since they usually have higher jitter.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/139
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/134
2019-08-04 19:34:11 +03:00
Aliaksandr Valialkin
1d4ddadbb1 app/vmselect/promql: further increase the allowed jitter for scrape interval
Real-world production data shows higher jitter than 1/8 of scrape interval.
This may results in gaps on the graph. So increase the allowed jitter to 1/4
of scrape interval in order to reduce the probability of gaps on the graphs
over time series with high jitter for scrape_interval.
2019-08-02 20:16:41 +03:00
Aliaksandr Valialkin
ade7bc30db app/vmselect/promql: tolerate higher jitter in scrape interval
Allow jitter for up to 1/8 instead of 1/16 for the scrape interval.
This should imrpove graphs when `step` is smaller than the `scrape_interval`.
2019-08-01 23:25:53 +03:00
Aliaksandr Valialkin
c994fbf500 app/vmselect/promql: add vm_slow_queries_total metric for counting slow queries
The query is slow if its execution time exceeds `-search.logSlowQueryDuration`
2019-07-31 03:36:45 +03:00
Aliaksandr Valialkin
071a122119 app/vmselect/promql: return NaN from histogram_quantile if at least a single bucket is broken 2019-07-31 01:18:11 +03:00
Aliaksandr Valialkin
b9a16b93e7 app/vmselect/promql: allow adjusting window for default rollup function
Default rollup function is `last_over_time`. It must support adjusting
the provided window in order to prevent from gaps on the graph
for window values smaller than scrape interval.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/134
2019-07-31 00:45:58 +03:00
Aliaksandr Valialkin
c901a6472f app/vmselect/promql: return NaN values if invalid bucket counts are passed to histogram_quantile
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/136
2019-07-30 22:05:55 +03:00
Aliaksandr Valialkin
aac482517f app/vmselect/promql: return NaN from count() over zero time series
This aligns `count` behavior with Prometheus.
2019-07-25 22:02:34 +03:00
Aliaksandr Valialkin
0e52357f35 app/vmselect/promql: properly calculate incremental aggregations grouped by __name__
Previously the following query may fail on multiple distinct metric names match:

    sum(count_over_time{__name__!=''}) by (__name__)
2019-07-25 21:53:26 +03:00
Aliaksandr Valialkin
092c9b39a8 app/vmselect/promql: remove empty time series after applying filters like q > 0
This should reduce CPU and RAM usage for queries over high number of time series.
2019-07-12 19:59:49 +03:00
Aliaksandr Valialkin
6875fb411a app/vmselect/promql: parallelize incremental aggregation to multiple CPU cores
This may reduce response times for aggregation over big number of time series
with small step between output data points.
2019-07-12 15:53:12 +03:00
Aliaksandr Valialkin
cbab86fd9d app/vmselect/promql: reduce RAM usage for aggregates over big number of time series
Calculate incremental aggregates for `aggr(metric_selector)` function instead of
keeping all the time series matching the given `metric_selector` in memory.
2019-07-10 13:03:36 +03:00
Aliaksandr Valialkin
ba8195c58e all: consistency renaming: bytesSize -> sizeBytes 2019-07-10 00:47:42 +03:00
Aliaksandr Valialkin
df6f17b82c app/vmselect/promql: mention -search.logSlowQueryDuration flag value in the slow query log message 2019-07-10 00:43:01 +03:00
Aliaksandr Valialkin
73ae889244 app/vmselect/promql: extract rmoeveGroupTags function for removing unneeded tags from MetricName according to the given modifierExpr 2019-07-09 23:20:58 +03:00
Aliaksandr Valialkin
603b34edbd app/vmselect/promql: properly preserve metric name after applying functions in any case from transformFuncsKeepMetricGroup 2019-07-09 23:10:49 +03:00
Aliaksandr Valialkin
bba07d05fe app/vmselect/promql: remove empty timeseries left after topk call 2019-07-04 19:43:07 +03:00
Aliaksandr Valialkin
858746fa6c app/vmselect/promql: gracefully handle duplicate timestamps in irate and rollup_rate funcs
Previously such timestamps result in `+Inf` results. Now the previous timestamp is used
for the calculations.
2019-07-03 12:41:30 +03:00
Aliaksandr Valialkin
a3abed80ff app/vmselect: do not return empty time series in /api/v1/query result 2019-07-01 17:16:26 +03:00
Aliaksandr Valialkin
c3c60bee45 app/vmselect: add -search.denyPartialResponse flag for disabling partial responses if some of vmstorage nodes are unavailable
Also accept `deny_partial_response` query arg in Prometheus API handlers. If it is set to true,
then return error if some of vmstorage nodes are unavailable.
2019-06-30 01:27:07 +03:00
Aliaksandr Valialkin
72a3050c41 app/vmselect/promql: consistency renaming: candlestick -> rollup_candlestick 2019-06-29 03:13:25 +03:00
Aliaksandr Valialkin
96342f1422 app/vmselect: add candlestick(m[d]) func for returning open, close, low and high rollups on the given time range d
This function is frequently used in financial apps. See https://en.wikipedia.org/wiki/Candlestick_chart
2019-06-27 18:46:54 +03:00
Aliaksandr Valialkin
4f54bcf90b app/vmselect/promql: suppress error when template func is used inside modifier list. Just leave it as is
Fixes https://github.com/VictoriaMetrics/VictoriaMetrics/issues/78
2019-06-25 20:43:57 +03:00
Aliaksandr Valialkin
d0bf4393a9 app/vmselect/promql: increase default value for -search.maxPointsPerTimeSeries from 10k to 30k
This may be required for subqueries with small steps. See https://github.com/VictoriaMetrics/VictoriaMetrics/issues/77
2019-06-24 22:53:25 +03:00
Aliaksandr Valialkin
334cf253c7 app/vmselect/promql: adjust value returned by linearRegression to the end of time range like Prometheus does
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/71
2019-06-24 22:46:03 +03:00
Aliaksandr Valialkin
14cd628948 app/vmselect/promql: add sum2 and sum2_over_time, geomean and geomean_over_time funcs.
These functions may be useful for statistic calculations.
2019-06-24 16:45:00 +03:00
Aliaksandr Valialkin
0eac538fc8 app/vmselect/promql: adjust the provided window only for range functions with dt in denominator
This should fix range function calculations such as `changes(m[d])` where `d` is smaller
than the scrape interval.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/72
2019-06-23 19:27:25 +03:00
Aliaksandr Valialkin
ec57e59154 app/vmselect/promql: use deriv_fast instead of deriv in ttf, since deriv calculations have been changed recently 2019-06-23 15:54:12 +03:00
Aliaksandr Valialkin
516062b162 app/vmselect/promql: adjust ttf calculation, so deriv(freev) for freev=m[d] could be properly calculated 2019-06-23 14:31:36 +03:00
Aliaksandr Valialkin
a4e040f5ef app/vmselect/promql: typo fixes in comments 2019-06-21 23:22:54 +03:00
Aliaksandr Valialkin
c05d443791 app/vmselect/promql: add deriv_fast function for calculating fast derivative
`deriv_fast` calculates derivative based on the first and the last point on the interval
instead of calculating linear regression based on all the data points on the interval.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/73
2019-06-21 23:05:48 +03:00
Aliaksandr Valialkin
98eafdbd58 app/vmselect/promql: use linear regression in deriv func like Prometheus does
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/73
2019-06-21 22:54:34 +03:00
Aliaksandr Valialkin
f334908c22 app/vmselect/promql: ajdust data model to the model used in Prometheus
Do not take into account data points on the range `[timestamp .. timestamp+step)`
when calculating value on the given `timestamp`.
Use only data points from the past when performing these calculations like Prometheus does.

This should reduce discrepancies between results returned by VictoriaMetrics
and results returned by Prometheus.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/72
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/71
2019-06-21 21:55:25 +03:00
Aliaksandr Valialkin
837e349b7d app/vmselect/promql: do not strip __name__ form time series after binary comparison operation
Example:

  foo > 10

Would leave `foo` name for all the matching time series on the left.
2019-06-21 13:08:02 +03:00
Aliaksandr Valialkin
e84b7641ef app/vmselect/promql: remove unused func keepLastValue; updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/69 2019-06-20 14:35:19 +03:00
Aliaksandr Valialkin
db042bf6d6 app/vmselect/promql: typo fix; updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/69 2019-06-20 14:33:52 +03:00
Aliaksandr Valialkin
947bc16f8c app/vmselect/promql: use dynamic limit on memory for concurrent queries 2019-06-12 23:18:23 +03:00
Aliaksandr Valialkin
8567e3463d app/vmselect/promql: merge non-overlapping duplicate time series in group_left and group_right joins 2019-06-12 20:33:01 +03:00
Aliaksandr Valialkin
88005237f4 app/vmselect/promql: swap binary operation with modifier in the error message for improved readability 2019-06-12 17:14:33 +03:00
Aliaksandr Valialkin
a71381ad2a app/vmselect/promql: list a sample of duplicate time series in the error message for group_left or group_right
This should improve troubleshooting for complex queries involving `group_left` and `group_right` modifiers.
2019-06-12 16:57:34 +03:00
Aliaksandr Valialkin
eea7da8e0c app/vmselect/promql: prevent from count_values explosion of timeseries, which could result in OOM 2019-06-11 01:03:18 +03:00
Aliaksandr Valialkin
e87a602209 app/vmselect/promql: skip superflouos timestamps copying in count_values 2019-06-11 00:44:09 +03:00
Aliaksandr Valialkin
ec84febc1c app/vmselect/promql: remove superflouos timeseries copy in histogram_quantile func 2019-06-11 00:39:35 +03:00
Aliaksandr Valialkin
1fab34fb5c app/vmselect/promql: remove superflouos timeseries copy in union func 2019-06-11 00:35:09 +03:00
Aliaksandr Valialkin
a6f368499d app/vmselect/promql: skip NaN values in count_values func 2019-06-10 22:42:41 +03:00
Aliaksandr Valialkin
913f888d0c app/vmselect/promql: properly handle {__name__ op "string"} queries
This has been broken in 7294ef333ad26f4f6578b783e97649e58b1f8945 .
2019-06-07 02:02:09 +03:00
Aliaksandr Valialkin
8f4790625d app/vmselect/promql: return the correct time series from quantile
Previously arbitrary time series could be returned from `quantile`
depending on sort order for the last data point in the selected range.

Fix this by returning the calculated time series.

Fixes https://github.com/VictoriaMetrics/VictoriaMetrics/issues/55
2019-06-06 17:33:53 +03:00
Aliaksandr Valialkin
2ff0d595b0 app/vmselect/promql: add -search.disableCache flag for disabling response caching
This may be useful for data back-filling, when the response caching
could interfere badly with newly added data points with timestamps
in the past.
2019-06-04 17:30:41 +03:00
Aliaksandr Valialkin
53242105fb app/vmselect/promql: allow escaping identifiers with \ and \xXX
Fixes https://github.com/VictoriaMetrics/VictoriaMetrics/issues/42
2019-05-31 17:35:54 +03:00
Aliaksandr Valialkin
a2c71f18a3 app/vmselect/promql: misspeling fix 2019-05-25 21:53:48 +03:00
Aliaksandr Valialkin
bdf696ef18 all: fix misspellings 2019-05-25 21:51:24 +03:00
Aliaksandr Valialkin
121a920a18 Makefile: add -s flag to go fmt in make fmt command 2019-05-25 21:44:36 +03:00
Aliaksandr Valialkin
2ff996e276 app/vmselect: log slow queries if their execution time exceeds -search.logSlowQueryDuration 2019-05-24 16:14:46 +03:00
Aliaksandr Valialkin
8e3eb5b39d app/vmselect/promql: add alias(q, name) function that sets the given name to all the time series in q 2019-05-24 02:42:10 +03:00
Aliaksandr Valialkin
bb048937bc app/vmselect/promql: add label_transform(q, label, regexp, replacement) function for replacing all the occurences of regexp with replacement in the given label for q 2019-05-23 16:26:07 +03:00
Aliaksandr Valialkin
24578b4bb1 all: open-sourcing cluster version 2019-05-23 00:25:38 +03:00
Aliaksandr Valialkin
1836c415e6 all: open-sourcing single-node version 2019-05-23 00:18:06 +03:00