feat(distributor): Add simulated latency #16733
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This change adds the ability to configure simulated latency on a per-tenant basis in the distributor. This is useful for testing and simulating real-world conditions. The implementation: - Adds a new `simulated_latency` field to the Limits struct using model.Duration - Adds a SimulatedLatency method to the Overrides struct - Adds the SimulatedLatency method to the Limits interface - Implements a waitSimulatedLatency method in the Distributor - Adds a defer call in the Push method to apply the latency at the end of processing The latency is applied at the end of request processing to ensure that all requests take at least the configured latency time, but without adding additional latency to requests that are already slow. When the configured latency is 0s, the feature is disabled.
github-actions
bot
added
the
type/docs
| @@ -485,11 +485,30 @@ func (p *pushTracker) doneWithResult(err error) { | |||
| } | |||
| } | |||
|
|
|||
| func (d *Distributor) waitSimulatedLatency(ctx context.Context, tenantID string, start time.Time) { | |||
| latency := d.validator.Limits.SimulatedPushLatency(tenantID) | |||
| if latency > 0 { | |||
There was a problem hiding this comment.
I believe this line can be removed as the condition is tested again when we check if wait > 0. For example, when latency is 0, then wait is negative.
However, I suppose someone could set latency to a negative number though? Is that a concern?
There was a problem hiding this comment.
Oh I see, you are using 0 as a special value to disable the feature.
| func (d *Distributor) Push(ctx context.Context, req *logproto.PushRequest) (*logproto.PushResponse, error) { | ||
| tenantID, err := tenant.TenantID(ctx) | ||
| if err != nil { | ||
| return nil, err | ||
| } | ||
| start := time.Now() | ||
| defer d.waitSimulatedLatency(ctx, tenantID, start) |
There was a problem hiding this comment.
Btw it's OK to use time.Now() as the argument to a deferred call. Go evaluates arguments for deferred functions when the defer is pushed onto the call stack, not when it is executed.
Suggested change
| defer d.waitSimulatedLatency(ctx, tenantID, start) | |
| defer d.waitSimulatedLatency(ctx, tenantID, time.Now()) |
grobinson-grafana
approved these changes
Mar 13, 2025
6 tasks
This was referenced Mar 20, 2025
This was referenced Mar 31, 2025
This was referenced Apr 14, 2025
This was referenced Apr 23, 2025
This was referenced May 5, 2025
This was referenced May 19, 2025
This was referenced May 26, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This change adds the ability to configure simulated latency on a per-tenant basis in the distributor. This is useful for testing and simulating real-world conditions.
The implementation:
simulated_latencyfield to the Limits struct using model.DurationThe latency is applied at the end of request processing to ensure that all requests take at least the configured latency time, but without adding additional latency to requests that are already slow.
When the configured latency is 0s, the feature is disabled.