Prometheus - v2.43.0

We are working on some performance improvements in Prometheus, which are only
built into Prometheus when compiling it using the Go tag stringlabels
(therefore they are not shipped in the default binaries). It uses a data
structure for labels that uses a single string to hold all the label/values,
resulting in a smaller heap size and some speedups in most cases. We would like
to encourage users who are interested in these improvements to help us measure
the gains on their production architecture. We are providing release artefacts
2.43.0+stringlabels and Docker images tagged v2.43.0-stringlabels with those
improvements for testing. #10991

• [FEATURE] Promtool: Add HTTP client configuration to query commands. #11487
• [FEATURE] Scrape: Add include_scrape_configs to include scrape configs from different files. #12019
• [FEATURE] HTTP client: Add no_proxy to exclude URLs from proxied requests. #12098
• [FEATURE] HTTP client: Add proxy_from_enviroment to read proxies from env variables. #12098
• [ENHANCEMENT] API: Add support for setting lookback delta per query via the API. #12088
• [ENHANCEMENT] API: Change HTTP status code from 503/422 to 499 if a request is canceled. #11897
• [ENHANCEMENT] Scrape: Allow exemplars for all metric types. #11984
• [ENHANCEMENT] TSDB: Add metrics for head chunks and WAL folders size. #12013
• [ENHANCEMENT] TSDB: Automatically remove incorrect snapshot with index that is ahead of WAL. #11859
• [ENHANCEMENT] TSDB: Improve Prometheus parser error outputs to be more comprehensible. #11682
• [ENHANCEMENT] UI: Scope group by labels to metric in autocompletion. #11914
• [BUGFIX] Scrape: Fix prometheus_target_scrape_pool_target_limit metric not set before reloading. #12002
• [BUGFIX] TSDB: Correctly update prometheus_tsdb_head_chunks_removed_total and prometheus_tsdb_head_chunks metrics when reading WAL. #11858
• [BUGFIX] TSDB: Use the correct unit (seconds) when recording out-of-order append deltas in the prometheus_tsdb_sample_ooo_delta metric. #12004