prometheus (2.4.3+ds-1) unstable; urgency=medium This release includes multiple bugfixes and features. Further, the WAL implementation has been re-written so the storage is not forwards compatible. Prometheus 2.3 storage will work on 2.4 but not vice-versa. -- Martín Ferrari Wed, 12 Sep 2018 15:19:58 +0000 prometheus (2.2.0+ds-1) unstable; urgency=medium NOTE: This release introduces improvements to the storage format and fixes a regression introduced in 2.1. As a result Prometheus servers upgraded to 2.2 cannot be downgraded to a lower version anymore! -- Martín Ferrari Wed, 14 Mar 2018 14:30:00 +0000 prometheus (2.0.0+ds3-1) unstable; urgency=medium BREAKING CHANGES The data format in Prometheus 2.0 has completely changed and is not backwards compatible with 1.x. Prometheus 2 will create a new data directory in /var/lib/prometheus/metrics2. The old data in /var/lib/prometheus/metrics can not be read by the new version. There is no conversion tool; if you want to retain access to the old data, it is necessary to run a separate Prometheus 1.x instance on that data directory. (This package makes no provision to allow this.) Command-line flags now take two slashes. For example `-web.listen-address` is now `--web.listen-address`. Rules are now stored in YAML files instead of using a domain-specific language. The `promtool update rules` command can assist you in the transition. -- Christoph Berg Fri, 15 Dec 2017 15:49:51 +0100 prometheus (1.5.0+ds-1) unstable; urgency=medium Differences from upstream: * The profile support for EC2 discovery has been disabled, as that requires a different version of the AWS SDK. * Kubernetes support has been completely removed, due to the insane amount of dependencies in their client library. -- Martín Ferrari Thu, 26 Jan 2017 17:39:25 +0000 prometheus (1.0.0+ds-1) unstable; urgency=medium First stable release! * [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 -- Martín Ferrari Tue, 19 Jul 2016 04:20:40 +0100 prometheus (0.20.0+ds-1) unstable; urgency=medium This release contains multiple breaking changes to the configuration schema. * [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 -- Martín Ferrari Fri, 24 Jun 2016 16:04:05 +0000 prometheus (0.19.2+ds-1) unstable; urgency=medium This version contains a breaking change to the query language. Please read the documentation on the changed 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 indivdual pages * [BUGFIX] Fix issue of hanging target scrapes -- Martín Ferrari Mon, 30 May 2016 09:52:19 +0000 prometheus (0.18.0+ds-1) unstable; urgency=medium * Support for the old alerting rule syntax was dropped for this version as announced with the 0.17.0 release. The new alerting rule syntax is documented [1] and a script to convert existing rules exists [2] [1] http://prometheus.io/docs/alerting/rules/ [2] https://gist.github.com/xbglowx/d798da98ff9937e33862b285d0121bde#gistcomment-1752515 * This release introduces a new chunk encoding with variable bit width, in parts inspired by Facebook's "Gorilla" paper. At the price of increased CPU cost and run-time of queries, the chunk space requirement drops by ~60% for typical data sets. The varbit encoding has to be enabled with the command line flag -storage.local.chunk-encoding-version=2, all newly created chunks will have the new chunk encoding. It is safe to use a mix of different chunk encodings. * Several storage improvements dramatically increase the performance of instant queries for the most recent value of time series, as it typically happens in rule evaluations and for federation. On servers with a lot of series churn (i.e. many older series that don't receive samples anymore), we have seen a 12x speedup of rule evaluations with CPU usage reduced by 75%. -- Martín Ferrari Tue, 19 Apr 2016 14:02:19 +0100 prometheus (0.16.2+ds-1) unstable; urgency=high Breaking changes in this release: * Remove undocumented 2nd argument of the `delta` function. * 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). -- Martín Ferrari Mon, 25 Jan 2016 07:34:22 -0300