github.com/yankunsam/loki/v2@v2.6.3-0.20220817130409-389df5235c27/docs/sources/operations/recording-rules.md (about) 1 --- 2 title: Recording Rules 3 --- 4 5 # Recording Rules 6 7 Recording rules are evaluated by the `ruler` component. Each `ruler` acts as its own `querier`, in the sense that it 8 executes queries against the store without using the `query-frontend` or `querier` components. It will respect all query 9 [limits](https://grafana.com/docs/loki/latest/configuration/#limits_config) put in place for the `querier`. 10 11 Loki's implementation of recording rules largely reuses Prometheus' code. 12 13 Samples generated by recording rules are sent to Prometheus using Prometheus' **remote-write** feature. 14 15 ## Write-Ahead Log (WAL) 16 17 All samples generated by recording rules are written to a WAL. The WAL's main benefit is that it persists the samples 18 generated by recording rules to disk, which means that if your `ruler` crashes, you won't lose any data. 19 We are trading off extra memory usage and slower start-up times for this functionality. 20 21 A WAL is created per tenant; this is done to prevent cross-tenant interactions. If all samples were to be written 22 to a single WAL, this would increase the chances that one tenant could cause data-loss for others. A typical scenario here 23 is that Prometheus will, for example, reject a remote-write request with 100 samples if just 1 of those samples is invalid in some way. 24 25 ### Start-up 26 27 When the `ruler` starts up, it will load the WALs for the tenants who have recording rules. These WAL files are stored 28 on disk and are loaded into memory. 29 30 Note: WALs are loaded one at a time upon start-up. This is a current limitation of the Loki ruler. 31 For this reason, it is adviseable that the number of rule groups serviced by a ruler be kept to a reasonable size, since 32 _no rule evaluation occurs while WAL replay is in progress (this includes alerting rules)_. 33 34 ### Truncation 35 36 WAL files are regularly truncated to reduce their size on disk. 37 [This guide](https://ganeshvernekar.com/blog/prometheus-tsdb-wal-and-checkpoint/#wal-truncation-and-checkpointing) 38 from one of the Prometheus maintainers (Ganesh Vernekar) gives an excellent overview of the truncation, checkpointing, 39 and replaying of the WAL. 40 41 ### Cleaner 42 43 <span style="background-color:#f3f973;">WAL Cleaner is an experimental feature.</span> 44 45 The WAL Cleaner watches for abandoned WALs (tenants who no longer have recording rules associated) and deletes them. 46 Enable this feature only if you are running into storage concerns with WALs that are too large. WALs should not grow 47 excessively large due to truncation. 48 49 ## Scaling 50 51 See Mimir's guide for [configuring Grafana Mimir hash rings](https://grafana.com/docs/mimir/latest/operators-guide/configuring/configuring-hash-rings/) for scaling the ruler using a ring. 52 53 Note: the `ruler` shards by rule _group_, not by individual rules. This is an artifact of the fact that Prometheus 54 recording rules need to run in order since one recording rule can reuse another - but this is not possible in Loki. 55 56 ## Deployment 57 58 The `ruler` needs to persist its WAL files to disk, and it incurs a bit of a start-up cost by reading these WALs into memory. 59 As such, it is recommended that you try to minimize churn of individual `ruler` instances since rule evaluation is blocked 60 while the WALs are being read from disk. 61 62 ### Kubernetes 63 64 It is recommended that you run the `rulers` using `StatefulSets`. The `ruler` will write its WAL files to persistent storage, 65 so a `Persistent Volume` should be utilised. 66 67 ## Remote-Write 68 69 ### Per-Tenant Limits 70 71 Remote-write can be configured at a global level in the base configuration, and certain parameters tuned specifically on 72 a per-tenant basis. Most of the configuration options [defined here](../../configuration/#ruler) 73 have [override options](../../configuration/#limits_config) (which can be also applied at runtime!). 74 75 ### Tuning 76 77 Remote-write can be tuned if the default configuration is insufficient (see [Failure Modes](#failure-modes) below). 78 79 There is a [guide](https://prometheus.io/docs/practices/remote_write/) on the Prometheus website, all of which applies to Loki, too. 80 81 ## Observability 82 83 Since Loki reuses the Prometheus code for recording rules and WALs, it also gains all of Prometheus' observability. 84 85 Prometheus exposes a number of metrics for its WAL implementation, and these have all been prefixed with `loki_ruler_wal_`. 86 87 For example: `prometheus_remote_storage_bytes_total` → `loki_ruler_wal_prometheus_remote_storage_bytes_total` 88 89 Additional metrics are exposed, also with the prefix `loki_ruler_wal_`. All per-tenant metrics contain a `tenant` 90 label, so be aware that cardinality could begin to be a concern if the number of tenants grows sufficiently large. 91 92 Some key metrics to note are: 93 - `loki_ruler_wal_appender_ready`: whether a WAL appender is ready to accept samples (1) or not (0) 94 - `loki_ruler_wal_prometheus_remote_storage_samples_total`: number of samples sent per tenant to remote storage 95 - `loki_ruler_wal_prometheus_remote_storage_samples...` 96 - `loki_ruler_wal_prometheus_remote_storage_samples_pending_total`: samples buffered in memory, waiting to be sent to remote storage 97 - `loki_ruler_wal_prometheus_remote_storage_samples_failed_total`: samples that failed when sent to remote storage 98 - `loki_ruler_wal_prometheus_remote_storage_samples_dropped_total`: samples dropped by relabel configurations 99 - `loki_ruler_wal_prometheus_remote_storage_samples_retried_total`: samples re-resent to remote storage 100 - `loki_ruler_wal_prometheus_remote_storage_highest_timestamp_in_seconds`: highest timestamp of sample appended to WAL 101 - `loki_ruler_wal_prometheus_remote_storage_queue_highest_sent_timestamp_seconds`: highest timestamp of sample sent to remote storage. 102 103 We've created a basic [dashboard in our loki-mixin](https://github.com/grafana/loki/tree/main/production/loki-mixin/dashboards/recording-rules.libsonnet) 104 which you can use to administer recording rules. 105 106 ## Failure Modes 107 108 ### Remote-Write Lagging 109 110 Remote-write can lag behind for many reasons: 111 112 1. Remote-write storage (Prometheus) is temporarily unavailable 113 2. A tenant is producing samples too quickly from a recording rule 114 3. Remote-write is tuned too low, creating backpressure 115 116 It can be determined by subtracting 117 `loki_ruler_wal_prometheus_remote_storage_queue_highest_sent_timestamp_seconds` from 118 `loki_ruler_wal_prometheus_remote_storage_highest_timestamp_in_seconds`. 119 120 In case 1, the `ruler` will continue to retry sending these samples until the remote storage becomes available again. Be 121 aware that if the remote storage is down for longer than `ruler.wal.max-age`, data loss may occur after truncation occurs. 122 123 In cases 2 & 3, you should consider [tuning](#tuning) remote-write appropriately. 124 125 Further reading: see [this blog post](https://grafana.com/blog/2021/04/12/how-to-troubleshoot-remote-write-issues-in-prometheus/) 126 by Prometheus maintainer Callum Styan. 127 128 ### Appender Not Ready 129 130 Each tenant's WAL has an "appender" internally; this appender is used to _append_ samples to the WAL. The appender is marked 131 as _not ready_ until the WAL replay is complete upon startup. If the WAL is corrupted for some reason, or is taking a long 132 time to replay, you can determine this by alerting on `loki_ruler_wal_appender_ready < 1`. 133 134 ### Corrupt WAL 135 136 If a disk fails or the `ruler` does not terminate correctly, there's a chance one or more tenant WALs can become corrupted. 137 A mechanism exists for automatically repairing the WAL, but this cannot handle every conceivable scenario. In this case, 138 the `loki_ruler_wal_corruptions_repair_failed_total` metric will be incremented. 139 140 ### Found another failure mode? 141 142 Please open an [issue](https://github.com/grafana/loki/issues) and tell us about it! 143