Added alert compliance test for Thanos #5315
There are no files selected for viewing
This file contains 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 file contains 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 file contains 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
| Original file line number | Diff line number | Diff line change |
|---|---|---|
|
|
@@ -4,6 +4,8 @@ | |
| package e2e_test | ||
|
|
||
| import ( | ||
| "bytes" | ||
| "fmt" | ||
| "io/ioutil" | ||
| "os" | ||
| "path/filepath" | ||
|
|
@@ -12,10 +14,16 @@ import ( | |
|
|
||
| "github.com/efficientgo/e2e" | ||
| e2edb "github.com/efficientgo/e2e/db" | ||
| common_cfg "github.com/prometheus/common/config" | ||
| "github.com/prometheus/prometheus/config" | ||
| "github.com/thanos-io/thanos/pkg/alert" | ||
| "github.com/thanos-io/thanos/pkg/httpconfig" | ||
| "github.com/thanos-io/thanos/pkg/testutil" | ||
| "github.com/thanos-io/thanos/test/e2e/e2ethanos" | ||
| ) | ||
|
|
||
| // TestPromQLCompliance tests PromQL compatibility against https://github.com/prometheus/compliance/tree/main/promql. | ||
| // NOTE: This requires dockerization of compliance framework: https://github.com/prometheus/compliance/pull/46 | ||
| // Test requires at least ~11m, so run this with `-test.timeout 9999m`. | ||
| func TestPromQLCompliance(t *testing.T) { | ||
| t.Skip("This is interactive test, it requires time to build up (scrape) the data. The data is also obtain from remote promlab servers.") | ||
|
|
@@ -24,7 +32,7 @@ func TestPromQLCompliance(t *testing.T) { | |
| testutil.Ok(t, err) | ||
| t.Cleanup(e.Close) | ||
|
|
||
| // Start receive + Querier. | ||
| receiverRunnable, err := e2ethanos.NewIngestingReceiver(e, "receive") | ||
| testutil.Ok(t, err) | ||
| queryReceive := e2edb.NewThanosQuerier(e, "query_receive", []string{receiverRunnable.InternalEndpoint("grpc")}) | ||
|
|
@@ -53,7 +61,7 @@ scrape_configs: | |
| )) | ||
| testutil.Ok(t, e2e.StartAndWaitReady(prom)) | ||
|
|
||
| // Start sidecar + Querier | ||
| sidecar := e2edb.NewThanosSidecar(e, "sidecar", prom, e2edb.WithImage("thanos")) | ||
| querySidecar := e2edb.NewThanosQuerier(e, "query_sidecar", []string{sidecar.InternalEndpoint("grpc")}, e2edb.WithImage("thanos")) | ||
| testutil.Ok(t, e2e.StartAndWaitReady(sidecar, querySidecar)) | ||
|
|
@@ -70,24 +78,28 @@ scrape_configs: | |
|
|
||
| t.Run("receive", func(t *testing.T) { | ||
| testutil.Ok(t, ioutil.WriteFile(filepath.Join(compliance.Dir(), "receive.yaml"), | ||
| []byte(promQLCompatConfig(prom, queryReceive, []string{"prometheus", "receive", "tenant_id"})), os.ModePerm)) | ||
|
|
||
| testutil.Ok(t, compliance.Exec(e2e.NewCommand( | ||
| "/promql-compliance-tester", | ||
| "-config-file", filepath.Join(compliance.InternalDir(), "receive.yaml"), | ||
| "-config-file", "/promql-test-queries.yml", | ||
| ))) | ||
| }) | ||
| t.Run("sidecar", func(t *testing.T) { | ||
| testutil.Ok(t, ioutil.WriteFile(filepath.Join(compliance.Dir(), "sidecar.yaml"), | ||
| []byte(promQLCompatConfig(prom, querySidecar, []string{"prometheus"})), os.ModePerm)) | ||
|
|
||
| testutil.Ok(t, compliance.Exec(e2e.NewCommand( | ||
| "/promql-compliance-tester", | ||
| "-config-file", filepath.Join(compliance.InternalDir(), "sidecar.yaml"), | ||
| "-config-file", "/promql-test-queries.yml", | ||
| ))) | ||
| }) | ||
| } | ||
|
|
||
| // nolint (it's still used in skipped test). | ||
| func promQLCompatConfig(reference *e2edb.Prometheus, target e2e.Runnable, dropLabels []string) string { | ||
| return `reference_target_config: | ||
| query_url: 'http://` + reference.InternalEndpoint("http") + `' | ||
|
|
||
|
|
@@ -105,3 +117,81 @@ query_tweaks: | |
| return ret | ||
| }() | ||
| } | ||
|
|
||
| // nolint (it's still used in skipped test). | ||
| type alwaysReadyProbe struct{} | ||
|
|
||
| // nolint (it's still used in skipped test). | ||
| func (p alwaysReadyProbe) Ready(e2e.Runnable) error { return nil } | ||
|
|
||
| // TestAlertCompliance tests Alert compatibility against https://github.com/prometheus/compliance/blob/main/alert_generator. | ||
| // NOTE: This requires dockerization of compliance framework: https://github.com/prometheus/compliance/pull/46 | ||
| func TestAlertCompliance_StatelessRuler(t *testing.T) { | ||
| t.Skip("This is an interactive test, using https://github.com/prometheus/compliance/tree/main/alert_generator testing is not optimized for CI runs") | ||
|
|
||
| e, err := e2e.NewDockerEnvironment("alert_compatibility") | ||
| testutil.Ok(t, err) | ||
| t.Cleanup(e.Close) | ||
|
|
||
| // Start receive + Querier. | ||
| receive, err := e2ethanos.NewIngestingReceiver(e, "receive") | ||
|
There was a problem hiding this comment. If we are using Thanos receiver in this case, the alerts will have the tenant label. If I remember correctly, the alerts compliance test requires the exact match of alerts labels. So maybe we need to configure the Querier to ignore the tenant label |
||
| testutil.Ok(t, err) | ||
| query := e2edb.NewThanosQuerier(e, "query_receive", []string{receive.InternalEndpoint("grpc")}) | ||
|
|
||
| compliance := e.Runnable("alert_generator_compliance_tester").WithPorts(map[string]int{"http": 8080}).Init(e2e.StartOptions{ | ||
| Image: "alert_generator_compliance_tester:latest", | ||
| // Batch job with HTTP port, we will start it with Exec, no need for readiness. | ||
| Readiness: alwaysReadyProbe{}, | ||
bwplotka marked this conversation as resolved.
Show resolved
Hide resolved
|
||
| Command: e2e.NewCommandWithoutEntrypoint("tail", "-f", "/dev/null"), | ||
| }) | ||
| ruler := e2ethanos.NewStatelessRuler(e, "1", "rules", []alert.AlertmanagerConfig{ | ||
| { | ||
| EndpointsConfig: httpconfig.EndpointsConfig{ | ||
| StaticAddresses: []string{compliance.InternalEndpoint("http")}, | ||
| Scheme: "http", | ||
| }, | ||
| Timeout: amTimeout, | ||
| APIVersion: alert.APIv1, | ||
| }, | ||
| }, []httpconfig.Config{ | ||
| { | ||
| EndpointsConfig: httpconfig.EndpointsConfig{ | ||
| StaticAddresses: []string{ | ||
| query.InternalEndpoint("http"), | ||
| }, | ||
| Scheme: "http", | ||
| }, | ||
| }, | ||
| }, []*config.RemoteWriteConfig{ | ||
| {URL: &common_cfg.URL{URL: urlParse(t, e2ethanos.RemoteWriteEndpoint(receive.InternalEndpoint("remote-write")))}, Name: "thanos-receiver"}, | ||
| }) | ||
| testutil.Ok(t, e2e.StartAndWaitReady(receive, query, ruler, compliance)) | ||
|
|
||
| // Pull rules.yaml: | ||
| { | ||
| var stdout bytes.Buffer | ||
| testutil.Ok(t, compliance.Exec(e2e.NewCommand("cat", "/rules.yaml"), e2e.WithExecOptionStdout(&stdout))) | ||
| testutil.Ok(t, os.MkdirAll(filepath.Join(ruler.Dir(), "rules"), os.ModePerm)) | ||
| testutil.Ok(t, os.WriteFile(filepath.Join(ruler.Dir(), "rules", "rules.yaml"), stdout.Bytes(), os.ModePerm)) | ||
| } | ||
|
|
||
| testutil.Ok(t, ioutil.WriteFile(filepath.Join(compliance.Dir(), "test-thanos.yaml"), | ||
| []byte(alertCompatConfig(receive, query)), os.ModePerm)) | ||
|
|
||
| fmt.Println(alertCompatConfig(receive, query)) | ||
|
|
||
| testutil.Ok(t, compliance.Exec(e2e.NewCommand( | ||
| "/alert_generator_compliance_tester", "-config-file", filepath.Join(compliance.InternalDir(), "test-thanos.yaml")), | ||
| )) | ||
| } | ||
|
|
||
| // nolint (it's still used in skipped test). | ||
| func alertCompatConfig(receive e2e.Runnable, query e2e.Runnable) string { | ||
| return `settings: | ||
| remote_write_url: '` + e2ethanos.RemoteWriteEndpoint(receive.InternalEndpoint("http")) + `' | ||
| query_base_url: 'http://` + query.InternalEndpoint("http") + `' | ||
| rules_and_alerts_api_base_url: 'http://` + query.InternalEndpoint("http") + `' | ||
|
There was a problem hiding this comment. I don't see stateless ruler being set as one endpoint of the Thanos Querier so I guess we cannot read anything from the Querier? Maybe we can use ruler's endpoint directly? |
||
| alert_reception_server_port: 8080 | ||
| alert_message_parser: default | ||
|
There was a problem hiding this comment. What about making this similar to https://github.com/thanos-io/thanos/blob/main/test/e2e/e2ethanos/services.go#L1058? There was a problem hiding this comment. I don't mind any, happy to change. |
||
| ` | ||
| } | ||
This file contains 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 file contains 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
Oops, something went wrong.
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.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Would be great to have another test case for TSDB ruler.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Same code is used, but yes - you played with it? Do you have it rdy? (we could add in another PR)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I don't have the code now. Happy to add it in another pr once this gets merged.