prometheus

The Prometheus monitoring system and time series database.

prometheus hacktoberfest metrics alerting graphing monitoring time-series

42 KiB

Raw Blame History

1.5.2 / 2017-02-10

[BUGFIX] Fix series corruption in a special case of series maintenance where the minimum series-file-shrink-ratio kicks in.
[BUGFIX] Fix two panic conditions both related to processing a series scheduled to be quarantined.
[ENHANCEMENT] Binaries built with Go1.7.5.

1.5.1 / 2017-02-07

[BUGFIX] Don't lose fully persisted memory series during checkpointing.
[BUGFIX] Fix intermittently failing relabeling.
[BUGFIX] Make -storage.local.series-file-shrink-ratio work.
[BUGFIX] Remove race condition from TestLoop.

1.5.0 / 2017-01-23

[CHANGE] Use lexicographic order to sort alerts by name.
[FEATURE] Add Joyent Triton discovery.
[FEATURE] Add scrape targets and alertmanager targets API.
[FEATURE] Add various persistence related metrics.
[FEATURE] Add various query engine related metrics.
[FEATURE] Add ability to limit scrape samples, and related metrics.
[FEATURE] Add labeldrop and labelkeep relabelling actions.
[FEATURE] Display current working directory on status-page.
[ENHANCEMENT] Strictly use ServiceAccount for in cluster configuration on Kubernetes.
[ENHANCEMENT] Various performance and memory-management improvements.
[BUGFIX] Fix basic auth for alertmanagers configured via flag.
[BUGFIX] Don't panic on decoding corrupt data.
[BUGFIX] Ignore dotfiles in data directory.
[BUGFIX] Abort on intermediate federation errors.

1.4.1 / 2016-11-28

[BUGFIX] Fix Consul service discovery

1.4.0 / 2016-11-25

[FEATURE] Allow configuring Alertmanagers via service discovery
[FEATURE] Display used Alertmanagers on runtime page in the UI
[FEATURE] Support profiles in AWS EC2 service discovery configuration
[ENHANCEMENT] Remove duplicated logging of Kubernetes client errors
[ENHANCEMENT] Add metrics about Kubernetes service discovery
[BUGFIX] Update alert annotations on re-evaluation
[BUGFIX] Fix export of group modifier in PromQL queries
[BUGFIX] Remove potential deadlocks in several service discovery implementations
[BUGFIX] Use proper float64 modulo in PromQL % binary operations
[BUGFIX] Fix crash bug in Kubernetes service discovery

1.3.1 / 2016-11-04

This bug-fix release pulls in the fixes from the 1.2.3 release.

[BUGFIX] Correctly handle empty Regex entry in relabel config.
[BUGFIX] MOD (%) operator doesn't panic with small floating point numbers.
[BUGFIX] Updated miekg/dns vendoring to pick up upstream bug fixes.
[ENHANCEMENT] Improved DNS error reporting.

1.2.3 / 2016-11-04

Note that this release is chronologically after 1.3.0.

[BUGFIX] Correctly handle end time before start time in range queries.
[BUGFIX] Error on negative -storage.staleness-delta
[BUGFIX] Correctly handle empty Regex entry in relabel config.
[BUGFIX] MOD (%) operator doesn't panic with small floating point numbers.
[BUGFIX] Updated miekg/dns vendoring to pick up upstream bug fixes.
[ENHANCEMENT] Improved DNS error reporting.

1.3.0 / 2016-11-01

This is a breaking change to the Kubernetes service discovery.

[CHANGE] Rework Kubernetes SD.
[FEATURE] Add support for interpolating target_label.
[FEATURE] Add GCE metadata as Prometheus meta labels.
[ENHANCEMENT] Add EC2 SD metrics.
[ENHANCEMENT] Add Azure SD metrics.
[ENHANCEMENT] Add fuzzy search to /graph textarea.
[ENHANCEMENT] Always show instance labels on target page.
[BUGFIX] Validate query end time is not before start time.
[BUGFIX] Error on negative -storage.staleness-delta

1.2.2 / 2016-10-30

[BUGFIX] Correctly handle on() in alerts.
[BUGFIX] UI: Deal properly with aborted requests.
[BUGFIX] UI: Decode URL query parameters properly.
[BUGFIX] Storage: Deal better with data corruption (non-monotonic timestamps).
[BUGFIX] Remote storage: Re-add accidentally removed timeout flag.
[BUGFIX] Updated a number of vendored packages to pick up upstream bug fixes.

1.2.1 / 2016-10-10

[BUGFIX] Count chunk evictions properly so that the server doesn't assume it runs out of memory and subsequencly throttles ingestion.
[BUGFIX] Use Go1.7.1 for prebuilt binaries to fix issues on MacOS Sierra.

1.2.0 / 2016-10-07

[FEATURE] Cleaner encoding of query parameters in /graph URLs.
[FEATURE] PromQL: Add minute() function.
[FEATURE] Add GCE service discovery.
[FEATURE] Allow any valid UTF-8 string as job name.
[FEATURE] Allow disabling local storage.
[FEATURE] EC2 service discovery: Expose ec2_instance_state.
[ENHANCEMENT] Various performance improvements in local storage.
[BUGFIX] Zookeeper service discovery: Remove deleted nodes.
[BUGFIX] Zookeeper service discovery: Resync state after Zookeeper failure.
[BUGFIX] Remove JSON from HTTP Accept header.
[BUGFIX] Fix flag validation of Alertmanager URL.
[BUGFIX] Fix race condition on shutdown.
[BUGFIX] Do not fail Consul discovery on Prometheus startup when Consul is down.
[BUGFIX] Handle NaN in changes() correctly.
[CHANGE] Experimental remote write path: Remove use of gRPC.
[CHANGE] Experimental remote write path: Configuration via config file rather than command line flags.
[FEATURE] Experimental remote write path: Add HTTP basic auth and TLS.
[FEATURE] Experimental remote write path: Support for relabelling.

1.1.3 / 2016-09-16

[ENHANCEMENT] Use golang-builder base image for tests in CircleCI.
[ENHANCEMENT] Added unit tests for federation.
[BUGFIX] Correctly de-dup metric families in federation output.

1.1.2 / 2016-09-08

[BUGFIX] Allow label names that coincide with keywords.

1.1.1 / 2016-09-07

[BUGFIX] Fix IPv6 escaping in service discovery integrations
[BUGFIX] Fix default scrape port assignment for IPv6

1.1.0 / 2016-09-03

[FEATURE] Add quantile and quantile_over_time.
[FEATURE] Add stddev_over_time and stdvar_over_time.
[FEATURE] Add various time and date functions.
[FEATURE] Added toUpper and toLower formatting to templates.
[FEATURE] Allow relabeling of alerts.
[FEATURE] Allow URLs in targets defined via a JSON file.
[FEATURE] Add idelta function.
[FEATURE] 'Remove graph' button on the /graph page.
[FEATURE] Kubernetes SD: Add node name and host IP to pod discovery.
[FEATURE] New remote storage write path. EXPERIMENTAL!
[ENHANCEMENT] Improve time-series index lookups.
[ENHANCEMENT] Forbid invalid relabel configurations.
[ENHANCEMENT] Improved various tests.
[ENHANCEMENT] Add crash recovery metric 'started_dirty'.
[ENHANCEMENT] Fix (and simplify) populating series iterators.
[ENHANCEMENT] Add job link on target page.
[ENHANCEMENT] Message on empty Alerts page.
[ENHANCEMENT] Various internal code refactorings and clean-ups.
[ENHANCEMENT] Various improvements in the build system.
[BUGFIX] Catch errors when unmarshalling delta/doubleDelta encoded chunks.
[BUGFIX] Fix data race in lexer and lexer test.
[BUGFIX] Trim stray whitespace from bearer token file.
[BUGFIX] Avoid divide-by-zero panic on query_range?step=0.
[BUGFIX] Detect invalid rule files at startup.
[BUGFIX] Fix counter reset treatment in PromQL.
[BUGFIX] Fix rule HTML escaping issues.
[BUGFIX] Remove internal labels from alerts sent to AM.

1.0.2 / 2016-08-24

[BUGFIX] Clean up old targets after config reload.

1.0.1 / 2016-07-21

[BUGFIX] Exit with error on non-flag command-line arguments.
[BUGFIX] Update example console templates to new HTTP API.
[BUGFIX] Re-add logging flags.

1.0.0 / 2016-07-18

[CHANGE] Remove deprecated query language keywords
[CHANGE] Change Kubernetes SD to require specifying Kubernetes role
[CHANGE] Use service address in Consul SD if available
[CHANGE] Standardize all Prometheus internal metrics to second units
[CHANGE] Remove unversioned legacy HTTP API
[CHANGE] Remove legacy ingestion of JSON metric format
[CHANGE] Remove deprecated target_groups configuration
[FEATURE] Add binary power operation to PromQL
[FEATURE] Add count_values aggregator
[FEATURE] Add -web.route-prefix flag
[FEATURE] Allow on(), by(), without() in PromQL with empty label sets
[ENHANCEMENT] Make topk/bottomk query functions aggregators
[BUGFIX] Fix annotations in alert rule printing
[BUGFIX] Expand alert templating at evaluation time
[BUGFIX] Fix edge case handling in crash recovery
[BUGFIX] Hide testing package flags from help output

0.20.0 / 2016-06-15

This release contains multiple breaking changes to the configuration schema.

[FEATURE] Allow configuring multiple Alertmanagers
[FEATURE] Add server name to TLS configuration
[FEATURE] Add labels for all node addresses and discover node port if available in Kubernetes SD
[ENHANCEMENT] More meaningful configuration errors
[ENHANCEMENT] Round scraping timestamps to milliseconds in web UI
[ENHANCEMENT] Make number of storage fingerprint locks configurable
[BUGFIX] Fix date parsing in console template graphs
[BUGFIX] Fix static console files in Docker images
[BUGFIX] Fix console JS XHR requests for IE11
[BUGFIX] Add missing path prefix in new status page
[CHANGE] Rename target_groups to static_configs in config files
[CHANGE] Rename names to files in file SD configuration
[CHANGE] Remove kubelet port config option in Kubernetes SD configuration

0.19.3 / 2016-06-14

[BUGFIX] Handle Marathon apps with zero ports
[BUGFIX] Fix startup panic in retrieval layer

0.19.2 / 2016-05-29

[BUGFIX] Correctly handle GROUP_LEFT and GROUP_RIGHT without labels in string representation of expressions and in rules.
[BUGFIX] Use -web.external-url for new status endpoints.

0.19.1 / 2016-05-25

[BUGFIX] Handle service discovery panic affecting Kubernetes SD
[BUGFIX] Fix web UI display issue in some browsers

0.19.0 / 2016-05-24

This version contains a breaking change to the query language. Please read the documentation on the grouping behavior of vector matching:

https://prometheus.io/docs/querying/operators/#vector-matching

[FEATURE] Add experimental Microsoft Azure service discovery
[FEATURE] Add ignoring modifier for binary operations
[FEATURE] Add pod discovery to Kubernetes service discovery
[CHANGE] Vector matching takes grouping labels from one-side
[ENHANCEMENT] Support time range on /api/v1/series endpoint
[ENHANCEMENT] Partition status page into individual pages
[BUGFIX] Fix issue of hanging target scrapes

0.18.0 / 2016-04-18

[BUGFIX] Fix operator precedence in PromQL
[BUGFIX] Never drop still open head chunk
[BUGFIX] Fix missing 'keep_common' when printing AST node
[CHANGE/BUGFIX] Target identity considers path and parameters additionally to host and port
[CHANGE] Rename metric prometheus_local_storage_invalid_preload_requests_total to prometheus_local_storage_non_existent_series_matches_total
[CHANGE] Support for old alerting rule syntax dropped
[FEATURE] Deduplicate targets within the same scrape job
[FEATURE] Add varbit chunk encoding (higher compression, more CPU usage – disabled by default)
[FEATURE] Add holt_winters query function
[FEATURE] Add relative complement unless operator to PromQL
[ENHANCEMENT] Quarantine series file if data corruption is encountered (instead of crashing)
[ENHANCEMENT] Validate Alertmanager URL
[ENHANCEMENT] Use UTC for build timestamp
[ENHANCEMENT] Improve index query performance (especially for active time series)
[ENHANCEMENT] Instrument configuration reload duration
[ENHANCEMENT] Instrument retrieval layer
[ENHANCEMENT] Add Go version to prometheus_build_info metric

0.17.0 / 2016-03-02

This version no longer works with Alertmanager 0.0.4 and earlier! The alerting rule syntax has changed as well but the old syntax is supported up until version 0.18.

All regular expressions in PromQL are anchored now, matching the behavior of regular expressions in config files.

[CHANGE] Integrate with Alertmanager 0.1.0 and higher
[CHANGE] Degraded storage mode renamed to rushed mode
[CHANGE] New alerting rule syntax
[CHANGE] Add label validation on ingestion
[CHANGE] Regular expression matchers in PromQL are anchored
[FEATURE] Add without aggregation modifier
[FEATURE] Send alert resolved notifications to Alertmanager
[FEATURE] Allow millisecond precision in configuration file
[FEATURE] Support AirBnB's Smartstack Nerve for service discovery
[ENHANCEMENT] Storage switches less often between regular and rushed mode.
[ENHANCEMENT] Storage switches into rushed mode if there are too many memory chunks.
[ENHANCEMENT] Added more storage instrumentation
[ENHANCEMENT] Improved instrumentation of notification handler
[BUGFIX] Do not count head chunks as chunks waiting for persistence
[BUGFIX] Handle OPTIONS HTTP requests to the API correctly
[BUGFIX] Parsing of ranges in PromQL fixed
[BUGFIX] Correctly validate URL flag parameters
[BUGFIX] Log argument parse errors
[BUGFIX] Properly handle creation of target with bad TLS config
[BUGFIX] Fix of checkpoint timing issue

0.16.2 / 2016-01-18

[FEATURE] Multiple authentication options for EC2 discovery added
[FEATURE] Several meta labels for EC2 discovery added
[FEATURE] Allow full URLs in static target groups (used e.g. by the blackbox_exporter)
[FEATURE] Add Graphite remote-storage integration
[FEATURE] Create separate Kubernetes targets for services and their endpoints
[FEATURE] Add clamp_{min,max} functions to PromQL
[FEATURE] Omitted time parameter in API query defaults to now
[ENHANCEMENT] Less frequent time series file truncation
[ENHANCEMENT] Instrument number of manually deleted time series
[ENHANCEMENT] Ignore lost+found directory during storage version detection
[CHANGE] Kubernetes masters renamed to api_servers
[CHANGE] "Healthy" and "unhealthy" targets are now called "up" and "down" in the web UI
[CHANGE] Remove undocumented 2nd argument of the delta function. (This is a BREAKING CHANGE for users of the undocumented 2nd argument.)
[BUGFIX] Return proper HTTP status codes on API errors
[BUGFIX] Fix Kubernetes authentication configuration
[BUGFIX] Fix stripped OFFSET from in rule evaluation and display
[BUGFIX] Do not crash on failing Consul SD initialization
[BUGFIX] Revert changes to metric auto-completion
[BUGFIX] Add config overflow validation for TLS configuration
[BUGFIX] Skip already watched Zookeeper nodes in serverset SD
[BUGFIX] Don't federate stale samples
[BUGFIX] Move NaN to end of result for topk/bottomk/sort/sort_desc/min/max
[BUGFIX] Limit extrapolation of delta/rate/increase
[BUGFIX] Fix unhandled error in rule evaluation

Some changes to the Kubernetes service discovery were integration since it was released as a beta feature.

0.16.1 / 2015-10-16

[FEATURE] Add irate() function.
[ENHANCEMENT] Improved auto-completion in expression browser.
[CHANGE] Kubernetes SD moves node label to instance label.
[BUGFIX] Escape regexes in console templates.

0.16.0 / 2015-10-09

BREAKING CHANGES:

Release tarballs now contain the built binaries in a nested directory.
The hash_mod relabeling action now uses MD5 hashes instead of FNV hashes to achieve a better distribution.
The DNS-SD meta label __meta_dns_srv_name was renamed to __meta_dns_name to reflect support for DNS record types other than SRV.
The default full refresh interval for the file-based service discovery has been increased from 30 seconds to 5 minutes.
In relabeling, parts of a source label that weren't matched by the specified regular expression are no longer included in the replacement output.
Queries no longer interpolate between two data points. Instead, the resulting value will always be the latest value before the evaluation query timestamp.
Regular expressions supplied via the configuration are now anchored to match full strings instead of substrings.
Global labels are not appended upon storing time series anymore. Instead, they are only appended when communicating with external systems (Alertmanager, remote storages, federation). They have thus also been renamed from global.labels to global.external_labels.
The names and units of metrics related to remote storage sample appends have been changed.
The experimental support for writing to InfluxDB has been updated to work with InfluxDB 0.9.x. 0.8.x versions of InfluxDB are not supported anymore.
Escape sequences in double- and single-quoted string literals in rules or query expressions are now interpreted like escape sequences in Go string literals (https://golang.org/ref/spec#String_literals).

Future breaking changes / deprecated features:

The delta() function had an undocumented optional second boolean argument to make it behave like increase(). This second argument will be removed in the future. Migrate any occurrences of delta(x, 1) to use increase(x) instead.
Support for filter operators between two scalar values (like 2 > 1) will be removed in the future. These will require a bool modifier on the operator, e.g. 2 > bool 1.

All changes:

[CHANGE] Renamed global.labels to global.external_labels.
[CHANGE] Vendoring is now done via govendor instead of godep.
[CHANGE] Change web UI root page to show the graphing interface instead of the server status page.
[CHANGE] Append global labels only when communicating with external systems instead of storing them locally.
[CHANGE] Change all regexes in the configuration to do full-string matches instead of substring matches.
[CHANGE] Remove interpolation of vector values in queries.
[CHANGE] For alert SUMMARY/DESCRIPTION template fields, cast the alert value to float64 to work with common templating functions.
[CHANGE] In relabeling, don't include unmatched source label parts in the replacement.
[CHANGE] Change default full refresh interval for the file-based service discovery from 30 seconds to 5 minutes.
[CHANGE] Rename the DNS-SD meta label __meta_dns_srv_name to __meta_dns_name to reflect support for other record types than SRV.
[CHANGE] Release tarballs now contain the binaries in a nested directory.
[CHANGE] Update InfluxDB write support to work with InfluxDB 0.9.x.
[FEATURE] Support full "Go-style" escape sequences in strings and add raw string literals.
[FEATURE] Add EC2 service discovery support.
[FEATURE] Allow configuring TLS options in scrape configurations.
[FEATURE] Add instrumentation around configuration reloads.
[FEATURE] Add bool modifier to comparison operators to enable boolean (0/1) output instead of filtering.
[FEATURE] In Zookeeper serverset discovery, provide __meta_serverset_shard label with the serverset shard number.
[FEATURE] Provide __meta_consul_service_id meta label in Consul service discovery.
[FEATURE] Allow scalar expressions in recording rules to enable use cases such as building constant metrics.
[FEATURE] Add label_replace() and vector() query language functions.
[FEATURE] In Consul service discovery, fill in the __meta_consul_dc datacenter label from the Consul agent when it's not set in the Consul SD config.
[FEATURE] Scrape all services upon empty services list in Consul service discovery.
[FEATURE] Add labelmap relabeling action to map a set of input labels to a set of output labels using regular expressions.
[FEATURE] Introduce __tmp as a relabeling label prefix that is guaranteed to not be used by Prometheus internally.
[FEATURE] Kubernetes-based service discovery.
[FEATURE] Marathon-based service discovery.
[FEATURE] Support multiple series names in console graphs JavaScript library.
[FEATURE] Allow reloading configuration via web handler at /-/reload.
[FEATURE] Updates to promtool to reflect new Prometheus configuration features.
[FEATURE] Add proxy_url parameter to scrape configurations to enable use of proxy servers.
[FEATURE] Add console templates for Prometheus itself.
[FEATURE] Allow relabeling the protocol scheme of targets.
[FEATURE] Add predict_linear() query language function.
[FEATURE] Support for authentication using bearer tokens, client certs, and CA certs.
[FEATURE] Implement unary expressions for vector types (-foo, +foo).
[FEATURE] Add console templates for the SNMP exporter.
[FEATURE] Make it possible to relabel target scrape query parameters.
[FEATURE] Add support for A and AAAA records in DNS service discovery.
[ENHANCEMENT] Fix several flaky tests.
[ENHANCEMENT] Switch to common routing package.
[ENHANCEMENT] Use more resilient metric decoder.
[ENHANCEMENT] Update vendored dependencies.
[ENHANCEMENT] Add compression to more HTTP handlers.
[ENHANCEMENT] Make -web.external-url flag help string more verbose.
[ENHANCEMENT] Improve metrics around remote storage queues.
[ENHANCEMENT] Use Go 1.5.1 instead of Go 1.4.2 in builds.
[ENHANCEMENT] Update the architecture diagram in the README.md.
[ENHANCEMENT] Time out sample appends in retrieval layer if the storage is backlogging.
[ENHANCEMENT] Make hash_mod relabeling action use MD5 instead of FNV to enable better hash distribution.
[ENHANCEMENT] Better tracking of targets between same service discovery mechanisms in one scrape configuration.
[ENHANCEMENT] Handle parser and query evaluation runtime panics more gracefully.
[ENHANCEMENT] Add IDs to H2 tags on status page to allow anchored linking.
[BUGFIX] Fix watching multiple paths with Zookeeper serverset discovery.
[BUGFIX] Fix high CPU usage on configuration reload.
[BUGFIX] Fix disappearing __params on configuration reload.
[BUGFIX] Make labelmap action available through configuration.
[BUGFIX] Fix direct access of protobuf fields.
[BUGFIX] Fix panic on Consul request error.
[BUGFIX] Redirect of graph endpoint for prefixed setups.
[BUGFIX] Fix series file deletion behavior when purging archived series.
[BUGFIX] Fix error checking and logging around checkpointing.
[BUGFIX] Fix map initialization in target manager.
[BUGFIX] Fix draining of file watcher events in file-based service discovery.
[BUGFIX] Add POST handler for /debug endpoints to fix CPU profiling.
[BUGFIX] Fix several flaky tests.
[BUGFIX] Fix busylooping in case a scrape configuration has no target providers defined.
[BUGFIX] Fix exit behavior of static target provider.
[BUGFIX] Fix configuration reloading loop upon shutdown.
[BUGFIX] Add missing check for nil expression in expression parser.
[BUGFIX] Fix error handling bug in test code.
[BUGFIX] Fix Consul port meta label.
[BUGFIX] Fix lexer bug that treated non-Latin Unicode digits as digits.
[CLEANUP] Remove obsolete federation example from console templates.
[CLEANUP] Remove duplicated Bootstrap JS inclusion on graph page.
[CLEANUP] Switch to common log package.
[CLEANUP] Update build environment scripts and Makefiles to work better with native Go build mechanisms and new Go 1.5 experimental vendoring support.
[CLEANUP] Remove logged notice about 0.14.x configuration file format change.
[CLEANUP] Move scrape-time metric label modification into SampleAppenders.
[CLEANUP] Switch from github.com/client_golang/model to github.com/common/model and related type cleanups.
[CLEANUP] Switch from github.com/client_golang/extraction to github.com/common/expfmt and related type cleanups.
[CLEANUP] Exit Prometheus when the web server encounters a startup error.
[CLEANUP] Remove non-functional alert-silencing links on alerting page.
[CLEANUP] General cleanups to comments and code, derived from golint, go vet, or otherwise.
[CLEANUP] When entering crash recovery, tell users how to cleanly shut down Prometheus.
[CLEANUP] Remove internal support for multi-statement queries in query engine.
[CLEANUP] Update AUTHORS.md.
[CLEANUP] Don't warn/increment metric upon encountering equal timestamps for the same series upon append.
[CLEANUP] Resolve relative paths during configuration loading.

0.15.1 / 2015-07-27

[BUGFIX] Fix vector matching behavior when there is a mix of equality and non-equality matchers in a vector selector and one matcher matches no series.
[ENHANCEMENT] Allow overriding GOARCH and GOOS in Makefile.INCLUDE.
[ENHANCEMENT] Update vendored dependencies.

0.15.0 / 2015-07-21