prometheus/docs/migration.md

title	sort_rank
Migration	10

Prometheus 3.0 migration guide

In line with our stability promise, the Prometheus 3.0 release contains a number of backwards incompatible changes. This document offers guidance on migrating from Prometheus 2.x to Prometheus 3.0 and newer versions.

Flags

• The following feature flags have been removed and they have been added to the default behavior of Prometheus v3:

  • promql-at-modifier
  • promql-negative-offset
  • remote-write-receiver
  • new-service-discovery-manager
  • expand-external-labels
    • Environment variable references ${var} or $var in external label values are replaced according to the values of the current environment variables.
    • References to undefined variables are replaced by the empty string. The $ character can be escaped by using $$.
  • no-default-scrape-port
    • Prometheus v3 will no longer add ports to scrape targets according to the specified scheme. Target will now appear in labels as configured.
    • If you rely on scrape targets like https://example.com/metrics or http://exmaple.com/metrics to be represented as https://example.com/metrics:443 and http://example.com/metrics:80 respectively, add them to your target URLs
  • agent
    • Instead use the dedicated --agent CLI flag.
  • auto-gomemlimit
    • Prometheus v3 will automatically set GOMEMLIMIT to match the Linux container memory limit. If there is no container limit, or the process is running outside of containers, the system memory total is used. To disable this, --no-auto-gomemlimit is available.
  • auto-gomaxprocs
    • Prometheus v3 will automatically set GOMAXPROCS to match the Linux container CPU quota. To disable this, --no-auto-gomaxprocs is available.

  Prometheus v3 will log a warning if you continue to pass these to --enable-feature.

Configuration

• The scrape job level configuration option scrape_classic_histograms has been renamed to always_scrape_classic_histograms. If you use the --enable-feature=native-histograms feature flag to ingest native histograms and you also want to ingest classic histograms that an endpoint might expose along with native histograms, be sure to add this configuration or change your configuration from the old name.
• The http_config.enable_http2 in remote_write items default has been changed to false. In Prometheus v2 the remote write http client would default to use http2. In order to parallelize multiple remote write queues across multiple sockets its preferable to not default to http2. If you prefer to use http2 for remote write you must now set http_config.enable_http2: true in your remote_write configuration section.

PromQL

Regular expressions match newlines

The . pattern in regular expressions in PromQL matches newline characters. With this change a regular expressions like .* matches strings that include \n. This applies to matchers in queries and relabel configs.

For example, the following regular expressions now match the accompanying strings, whereas in Prometheus v2 these combinations didn't match. - .* additionally matches foo\n and Foo\nBar - foo.?bar additionally matches foo\nbar - foo.+bar additionally matches foo\nbar

If you want Prometheus v3 to behave like v2, you will have to change your regular expressions by replacing all . patterns with [^\n], e.g. foo[^\n]*.

Range selectors and lookback exclude samples coinciding with the left boundary

Lookback and range selectors are now left-open and right-closed (previously left-closed and right-closed), which makes their behavior more consistent. This change affects queries where the left boundary of a range or the lookback delta coincides with the timestamp of one or more samples.

For example, assume we are querying a timeseries with evenly spaced samples exactly 1 minute apart. Before Prometheus v3, a range query with 5m would usually return 5 samples. But if the query evaluation aligns perfectly with a scrape, it would return 6 samples. In Prometheus v3 queries like this will always return 5 samples given even spacing.

This change will typically affect subqueries because their evaluation timing is naturally perfectly evenly spaced and aligned with timestamps that are multiples of the subquery resolution. Furthermore, query frontends often align subqueries to multiples of the step size. In combination, this easily creates a situation of perfect mutual alignment, often unintended and unknown by the user, so that the new behavior might come as a surprise. Before Prometheus V3, a subquery of foo[1m:1m] on such a system might have always returned two points, allowing for rate calculations. In Prometheus V3, however, such a subquery will only return one point, which is insufficient for a rate or increase calculation, resulting in No Data returned.

Such queries will need to be rewritten to extend the window to properly cover more than one point. In this example, foo[2m:1m] would always return two points no matter the query alignment. The exact form of the rewritten query may depend on the intended results and there is no universal drop-in replacement for queries whose behavior has changed.

Tests are similarly more likely to affected. To fix those either adjust the expected number of samples or extend the range.

holt_winters function renamed

The holt_winters function has been renamed to double_exponential_smoothing and is now guarded by the promql-experimental-functions feature flag. If you want to keep using holt_winters, you have to do both of these things:

• Rename holt_winters to double_exponential_smoothing in your queries.
• Pass --enable-feature=promql-experimental-functions in your Prometheus CLI invocation.

Scrape protocols

Prometheus v3 is more strict concerning the Content-Type header received when scraping. Prometheus v2 would default to the standard Prometheus text protocol if the target being scraped did not specify a Content-Type header or if the header was unparsable or unrecognised. This could lead to incorrect data being parsed in the scrape. Prometheus v3 will now fail the scrape in such cases.

If a scrape target is not providing the correct Content-Type header the fallback protocol can be specified using the fallback_scrape_protocol parameter. See Prometheus scrape_config documentation.

This is a breaking change as scrapes that may have succeeded with Prometheus v2 may now fail if this fallback protocol is not specified.

Miscellaneous

TSDB format and downgrade

The TSDB format has been changed slightly in Prometheus v2.55 in preparation for changes to the index format. Consequently, a Prometheus v3 TSDB can only be read by a Prometheus v2.55 or newer. Keep that in mind when upgrading to v3 -- you will be only able to downgrade to v2.55, not lower, without losing your TSDB persitent data.

As an extra safety measure, you could optionally consider upgrading to v2.55 first and confirm Prometheus works as expected, before upgrading to v3.

TSDB storage contract

TSDB compatible storage is now expected to return results matching the specified selectors. This might impact some third party implementations, most likely implementing remote_read.

This contract is not explicitly enforced, but can cause undefined behavior.

UTF-8 names

Prometheus v3 supports UTF-8 in metric and label names. This means metric and label names can change after upgrading according to what is exposed by endpoints. Furthermore, metric and label names that would have previously been flagged as invalid no longer will be.

Users wishing to preserve the original validation behavior can update their Prometheus yaml configuration to specify the legacy validation scheme:

global:
  metric_name_validation_scheme: legacy

Or on a per-scrape basis:

scrape_configs:
  - job_name: job1
    metric_name_validation_scheme: utf8
  - job_name: job2
    metric_name_validation_scheme: legacy

Log message format

Prometheus v3 has adopted log/slog over the previous go-kit/log. This results in a change of log message format. An example of the old log format is:

ts=2024-10-23T22:01:06.074Z caller=main.go:627 level=info msg="No time or size retention was set so using the default time retention" duration=15d
ts=2024-10-23T22:01:06.074Z caller=main.go:671 level=info msg="Starting Prometheus Server" mode=server version="(version=, branch=, revision=91d80252c3e528728b0f88d254dd720f6be07cb8-modified)"
ts=2024-10-23T22:01:06.074Z caller=main.go:676 level=info build_context="(go=go1.23.0, platform=linux/amd64, user=, date=, tags=unknown)"
ts=2024-10-23T22:01:06.074Z caller=main.go:677 level=info host_details="(Linux 5.15.0-124-generic #134-Ubuntu SMP Fri Sep 27 20:20:17 UTC 2024 x86_64 gigafips (none))"

a similar sequence in the new log format looks like this:

time=2024-10-24T00:03:07.542+02:00 level=INFO source=/home/user/go/src/github.com/prometheus/prometheus/cmd/prometheus/main.go:640 msg="No time or size retention was set so using the default time retention" duration=15d
time=2024-10-24T00:03:07.542+02:00 level=INFO source=/home/user/go/src/github.com/prometheus/prometheus/cmd/prometheus/main.go:681 msg="Starting Prometheus Server" mode=server version="(version=, branch=, revision=7c7116fea8343795cae6da42960cacd0207a2af8)"
time=2024-10-24T00:03:07.542+02:00 level=INFO source=/home/user/go/src/github.com/prometheus/prometheus/cmd/prometheus/main.go:686 msg="operational information" build_context="(go=go1.23.0, platform=linux/amd64, user=, date=, tags=unknown)" host_details="(Linux 5.15.0-124-generic #134-Ubuntu SMP Fri Sep 27 20:20:17 UTC 2024 x86_64 gigafips (none))" fd_limits="(soft=1048576, hard=1048576)" vm_limits="(soft=unlimited, hard=unlimited)"

le and quantile label values

In Prometheus v3, the values of the le label of classic histograms and the quantile label of summaries are normalized upon ingestion. In Prometheus v2 the value of these labels depended on the scrape protocol (protobuf vs text format) in some situations. This led to label values changing based on the scrape protocol. E.g. a metric exposed as my_classic_hist{le="1"} would be ingested as my_classic_hist{le="1"} via the text format, but as my_classic_hist{le="1.0"} via protobuf. This changed the identity of the metric and caused problems when querying the metric. In Prometheus v3 these label values will always be normalized to a float like representation. I.e. the above example will always result in my_classic_hist{le="1.0"} being ingested into prometheus, no matter via which protocol. The effect of this change is that alerts, recording rules and dashboards that directly reference label values as whole numbers such as le="1" will stop working.

Ways to deal with this change either globally or on a per metric basis:

• Fix references to integer le, quantile label values, but otherwise do nothing and accept that some queries that span the transition time will produce inaccurate or unexpected results. This is the recommended solution.
• Use metric_relabel_config to retain the old labels when scraping targets. This should only be applied to metrics that currently produce such labels.

    metric_relabel_configs:
      - source_labels:
          - quantile
        target_label: quantile
        regex: (\d+)\.0+
      - source_labels:
          - le
          - __name__
        target_label: le
        regex: (\d+)\.0+;.*_bucket

Disallow configuring Alertmanager with the v1 API

Prometheus 3 no longer supports Alertmanager's v1 API. Effectively Prometheus 3 requires Alertmanager 0.16.0 or later. Users with older Alertmanager versions or configurations that use alerting: alertmanagers: [api_version: v1] need to upgrade Alertmanager and change their configuration to use api_version: v2.

Prometheus 2.0 migration guide

For the Prometheus 1.8 to 2.0 please refer to the Prometheus v2.55 documentation.