mirrors/external-dns
1
0
Fork 0
mirror of https://github.com/kubernetes-sigs/external-dns.git synced 2025-08-06 09:36:58 +02:00
Code Issues Packages Projects Releases Wiki Activity

Apply suggestions from code review

Browse Source 
Co-authored-by: Michel Loiseleur <97035654+mloiseleur@users.noreply.github.com>
This commit is contained in:
Thibault Jamet 2024-08-13 11:16:28 +02:00
parent 173604854d
commit ef0dd29cf5
No known key found for this signature in database
GPG Key ID: 9D28A304A3810C17
1 changed files with 23 additions and 23 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

46
docs/rate-limits.md
View File

@ -6,10 +6,10 @@ DNS provider API rate limits considerations
By design, external-dns refreshes all the records of a zone using API calls. By design, external-dns refreshes all the records of a zone using API calls.
This refresh may happen peridically and upon any changed object if the flag `--events` is enabled. This refresh may happen peridically and upon any changed object if the flag `--events` is enabled.
Depending on the size of the zone and the infrastructure deployment, this may lead to having external-dns Depending on the size of the zone and the infrastructure deployment, this may lead to external-dns
hitting the DNS provider's rate-limits more easily. hitting the DNS provider's rate-limits more easily.
In particular, it has been found that, with 200k records in an AWS Route53 zone, each refresh triggers around In particular, it has been found that with 200k records in an AWS Route53 zone, each refresh triggers around
70 API calls to retrieve all the records, making it more likely to hit the AWS Route53 API rate limits. 70 API calls to retrieve all the records, making it more likely to hit the AWS Route53 API rate limits.
To prevent this problem from happening, external-dns has implemented a cache to reduce the pressure on the DNS To prevent this problem from happening, external-dns has implemented a cache to reduce the pressure on the DNS
@ -24,19 +24,19 @@ The major trade-off of this setting relies in the ability to recover from a dele
As the DNS records are cached in memory, external-dns will not be made aware of the missing records and will hence As the DNS records are cached in memory, external-dns will not be made aware of the missing records and will hence
take a longer time to restore the deleted or modified record on the provider side. take a longer time to restore the deleted or modified record on the provider side.
This option is enabled using the `--provider-cache-time=15m` command line argument, and disabled when `--provider-cache-time=0m` This option is enabled using the `--provider-cache-time=15m` command line argument, and turned off when `--provider-cache-time=0m`
## Monitoring ## Monitoring
You can evaluate the behaviour of the cache thanks to the built-in metrics You can evaluate the behaviour of the cache thanks to the built-in metrics
* `external_dns_provider_cache_records_calls` * `external_dns_provider_cache_records_calls`
The number of calls to the provider cache Records list. * The number of calls to the provider cache Records list.
The label `from_cache=true` indicates that the records were retrieved from memory and the DNS provider was not reached * The label `from_cache=true` indicates that the records were retrieved from memory and the DNS provider was not reached
The label `from_cache=false` indicates that the cache was not used and the records were retrieved from the provider * The label `from_cache=false` indicates that the cache was not used and the records were retrieved from the provider
* `external_dns_provider_cache_apply_changes_calls` * `external_dns_provider_cache_apply_changes_calls`
The number of calls to the provider cache ApplyChanges. * The number of calls to the provider cache ApplyChanges.
Each ApplyChange systematically invalidates the cache and makes subsequent Records list to be retrieved from the provider without cache. * Each ApplyChange systematically invalidates the cache and makes subsequent Records list to be retrieved from the provider without cache.
## Related options ## Related options
@ -45,26 +45,26 @@ or provider-specific options to fine-tune the behaviour of external-dns
to match the specific needs of your deployments, with the goal to reduce the number of API calls to your DNS provider. to match the specific needs of your deployments, with the goal to reduce the number of API calls to your DNS provider.
* Google * Google
`--google-batch-change-interval=1s` When using the Google provider, set the interval between batch changes. ($EXTERNAL_DNS_GOOGLE_BATCH_CHANGE_INTERVAL) * `--google-batch-change-interval=1s` When using the Google provider, set the interval between batch changes. ($EXTERNAL_DNS_GOOGLE_BATCH_CHANGE_INTERVAL)
`--google-batch-change-size=1000` When using the Google provider, set the maximum number of changes that will be applied in each batch. * `--google-batch-change-size=1000` When using the Google provider, set the maximum number of changes that will be applied in each batch.
* AWS * AWS
`--aws-batch-change-interval=1s` When using the AWS provider, set the interval between batch changes. * `--aws-batch-change-interval=1s` When using the AWS provider, set the interval between batch changes.
`--aws-batch-change-size=1000` When using the AWS provider, set the maximum number of changes that will be applied in each batch. * `--aws-batch-change-size=1000` When using the AWS provider, set the maximum number of changes that will be applied in each batch.
`--aws-batch-change-size-bytes=32000` When using the AWS provider, set the maximum byte size that will be applied in each batch. * `--aws-batch-change-size-bytes=32000` When using the AWS provider, set the maximum byte size that will be applied in each batch.
`--aws-batch-change-size-values=1000` When using the AWS provider, set the maximum total record values that will be applied in each batch. * `--aws-batch-change-size-values=1000` When using the AWS provider, set the maximum total record values that will be applied in each batch.
`--aws-zones-cache-duration=0s` When using the AWS provider, set the zones list cache TTL (0s to disable). * `--aws-zones-cache-duration=0s` When using the AWS provider, set the zones list cache TTL (0s to disable).
`--[no-]aws-zone-match-parent` Expand limit possible target by sub-domains * `--[no-]aws-zone-match-parent` Expand limit possible target by sub-domains
* Cloudflare * Cloudflare
`--cloudflare-dns-records-per-page=100` When using the Cloudflare provider, specify how many DNS records listed per page, max possible 5,000 (default: 100) * `--cloudflare-dns-records-per-page=100` When using the Cloudflare provider, specify how many DNS records listed per page, max possible 5,000 (default: 100)
* OVH * OVH
`--ovh-api-rate-limit=20` When using the OVH provider, specify the API request rate limit, X operations by seconds (default: 20) * `--ovh-api-rate-limit=20` When using the OVH provider, specify the API request rate limit, X operations by seconds (default: 20)
* Global * Global
`--registry=txt` The registry implementation to use to keep track of DNS record ownership (default: txt, options: txt, noop, dynamodb, aws-sd) * `--registry=txt` The registry implementation to use to keep track of DNS record ownership (default: txt, options: txt, noop, dynamodb, aws-sd)
`--txt-cache-interval=0s` The interval between cache synchronizations in duration format (default: disabled) * `--txt-cache-interval=0s` The interval between cache synchronizations in duration format (default: disabled)
`--interval=1m0s` The interval between two consecutive synchronizations in duration format (default: 1m) * `--interval=1m0s` The interval between two consecutive synchronizations in duration format (default: 1m)
`--min-event-sync-interval=5s` The minimum interval between two consecutive synchronizations triggered from kubernetes events in duration format (default: 5s) * `--min-event-sync-interval=5s` The minimum interval between two consecutive synchronizations triggered from kubernetes events in duration format (default: 5s)
`--[no-]events` When enabled, in addition to running every interval, the reconciliation loop will get triggered when supported sources change (default: disabled) * `--[no-]events` When enabled, in addition to running every interval, the reconciliation loop will get triggered when supported sources change (default: disabled)
A general recommendation is to enable `--events` and keep `--min-event-sync-interval` relatively low to have a better responsiveness when records are A general recommendation is to enable `--events` and keep `--min-event-sync-interval` relatively low to have a better responsiveness when records are
created or updated inside the cluster. created or updated inside the cluster.