Add Managed Connection Pooling support for an AlloyDB Instance #5611

nancynh · 2025-11-11T23:32:36Z

BRIEF Change description

Fixes #5610. We add a new nested object field connection_pool_config to the AlloyDB instance. This will allow customers to enable managed connection pooling and set various settings on it.

WHY do we need this change?

Special notes for your reviewer:

I did re-generate the MockGCP files to be on an newer release version of the googeapis. I did this only for the AlloyDB MockGCP files though.

I also did manually edit some files like the mapper, fuzzer, controller, etc because the commands to generate the changes didn't work for me.

Does this PR add something which needs to be 'release noted'?

alloydb: added `connection_pool_config` to the AlloyDB Instance.

Reviewer reviewed release note.

Additional documentation e.g., references, usage docs, etc.:

Intended Milestone

Please indicate the intended milestone.

Reviewer tagged PR with the actual milestone.

Tests you have done

Run make ready-pr to ensure this PR is ready for review.
Perform necessary E2E testing for changed resources.

nancynh · 2025-11-12T00:37:09Z

The fuzz-roundtrippers check is failing due to a diff that is unrelated to my changes:

        fuzzkrm.go:230: roundtrip failed for KRM *v1beta1.AlloyDBInstanceSpec; diff:
              (*alloydbpb.Instance)(Inverse(protocmp.Transform, protocmp.Message{
              	"@type":             s"google.cloud.alloydb.v1beta.Instance",
            - 	"activation_policy": s"NEVER",
              	"annotations":       map[string]string{"13d9dc7ff603b0ed": "1a392cdbc4af8f81", "2c756d6e2ea4d678": "77967f5415fd5ffb", "6ac38872617a228f": "4bbea4a394aeace9"},
              }))

I think that activation policy field is being worked on separately in #5052

maqiuyujoyce

The fuzz-roundtrippers check is failing due to a diff that is unrelated to my changes: ...

Could you try to mark this field as unimplemented in the fuzzer? Fuzzer test is a blocker for any PR to be merged.

maqiuyujoyce · 2025-11-15T02:30:39Z

mockgcp/generated/mockgcp/cloud/alloydb/v1beta/resources.pb.go

@@ -172,6 +172,8 @@ const (
 	DatabaseVersion_POSTGRES_15 DatabaseVersion = 3
 	// The database version is Postgres 16.
 	DatabaseVersion_POSTGRES_16 DatabaseVersion = 4
+	// The database version is Postgres 17.


Just curious, how did you update these files? Asking because git.versions file is not updated but the .pb.go files are.

I did change the git.versions file locally and ran ./dev/tasks/update-mockgcp. I just only kept the files that were updated in the AlloyDB folder cause I felt like updating the other API's mockgcp files was out of scope of this PR. But happy to send out a separate PR if need be to update the mockgcp for all APIs

If so, then yes, you'll want to update all APIs, otherwise others may revert your change here.

Hmm, I'm having some issues while updating the mockgcp to the latest SHA. After updating them and running make ready-pr I get this error:

level=error msg="Running error: can't run linter goanalysis_metalinter\nfindcall: failed to load package dataplexpb: could not load export data: no export data for \"github.com/GoogleCloudPlatform/k8s-config-connector/mockgcp/generated/google/cloud/dataplex/v1\""

I did make sure that I had go setup correctly (re-did the go setup as per the guide), cleared the golangci-lint cache, tried to force mock gcp to use a different SHA by hard-coding it in dev/tasks/update-mockgcp, and also tried to do the same on a new clean branch synced to the latest off of the master branch.

Plus it seems like updating them also ends up breaking a bunch of tests. I've already reverted the file changes I did for mockgcp. You can see the test failures at: https://github.com/GoogleCloudPlatform/k8s-config-connector/actions/runs/19944914990/job/57191832228?pr=5611

I've also asked about this in the internal chat

Yes, we can follow up separately. Thanks for working on this!

maqiuyujoyce · 2025-11-15T02:32:01Z

mockgcp/generated/mockgcp/cloud/alloydb/v1beta/resources.pb.go

@@ -5316,6 +5307,8 @@ type Instance_ConnectionPoolConfig struct {
 	Enabled bool `protobuf:"varint,12,opt,name=enabled,proto3" json:"enabled,omitempty"`
 	// Optional. Connection Pool flags, as a list of "key": "value" pairs.
 	Flags map[string]string `protobuf:"bytes,13,rep,name=flags,proto3" json:"flags,omitempty" protobuf_key:"bytes,1,opt,name=key,proto3" protobuf_val:"bytes,2,opt,name=value,proto3"`
+	// Output only. The number of running poolers per instance.


This field is added to Instance_ConnectionPoolConfig type, but it is not reflected in instance type (instance_types.go or types.generated.go). Looks like something is missing here?

Yeah, when I generated the changes for instance_types.go and types.generated.go, the connection_pool_config.pooler_count field wasn't included. I've added it in manually in types.generated.go now (after running make ready-pr this also updated some other files too)

For instance_types.go, should be the field be under AlloyDBInstanceStatus? Similar to other output only fields like OutboundPublicIPAddresses. The only difference is though is that pooler_count isn't a top-level field like those are...

I've added it in manually in types.generated.go now (after running make ready-pr this also updated some other files too)

Manual change should be done in XX_types.go file only. So you'll want to move the entire struct to instance_types.go file with your manual change. Otherwise your manual change in the generated file may be overridden.

Besides, could you file a GitHub issue about this and mention me there? It's unexpected to exclude fields in the types generation.

For instance_types.go, should be the field be under AlloyDBInstanceStatus? Similar to other output only fields like OutboundPublicIPAddresses. The only difference is though is that pooler_count isn't a top-level field like those are...

It should be under AlloyDBInstanceObservedState struct. And you can keep the original path under it. To summarize, you should make the following changes:

In types.generated.go: Remove struct Instance_ConnectionPoolConfig.

In instance_types.go: Add two structs, one is configurable, one is output only (with ObservedState suffix).

// +kcc:proto=google.cloud.alloydb.v1beta.Instance.ConnectionPoolConfig type Instance_ConnectionPoolConfig struct { // Optional. Whether to enable Managed Connection Pool (MCP). // +kcc:proto:field=google.cloud.alloydb.v1beta.Instance.ConnectionPoolConfig.enabled Enabled *bool `json:"enabled,omitempty"` // Optional. Connection Pool flags, as a list of "key": "value" pairs. // +kcc:proto:field=google.cloud.alloydb.v1beta.Instance.ConnectionPoolConfig.flags Flags map[string]string `json:"flags,omitempty"` } // +kcc:proto=google.cloud.alloydb.v1beta.Instance.ConnectionPoolConfig type Instance_ConnectionPoolConfigObservedState struct { // Output only. The number of running poolers per instance. // +kcc:proto:field=google.cloud.alloydb.v1beta.Instance.ConnectionPoolConfig.pooler_count PoolerCount *int32 `json:"pooler_count,omitempty"` }

Also in instance_types.go: Support AlloyDBInstanceObservedState.

Uncomment the two places where AlloyDBInstanceObservedState is commented out.

Add the following field in struct:

// Output for Managed Connection Pool (MCP). ConnectionPoolConfig *Instance_ConnectionPoolConfigObservedState `json:"connectionPoolConfig,omitempty"`

Run make ready-pr.

Thanks for the very detailed explanation+steps! I filed the bug at #5753, and have made the necessary changes

Thank you, Nancy!

After looking at the issue description, I suspect it is the problem of the tool mentioned in the doc, which hasn't been updated for a while (and I actually never used it). We probably want to update our doc.

^ This is outside of the scope of this PR though.

...testdata/basic/alloydb/v1beta1/alloydbinstance/alloydbinstanceconnectionpoolconfig/_http.log

google-oss-prow · 2025-11-18T23:41:04Z

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by:
Once this PR has been reviewed and has the lgtm label, please ask for approval from maqiuyujoyce. For more information see the Kubernetes Code Review Process.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

nancynh · 2025-11-21T22:34:42Z

The fuzz-roundtrippers check is failing due to a diff that is unrelated to my changes: ...

Could you try to mark this field as unimplemented in the fuzzer? Fuzzer test is a blocker for any PR to be merged.

I think I fixed the fuzzer tests now? I also excluded a few other AlloyDB fields that haven't been onboarded yet to KCC

…and under spec for the AlloyDB instance fuzzer

…led is true

…oolConfig

nancynh · 2025-12-08T20:48:23Z

@maqiuyujoyce the PR is ready for your review again - I did have some issues on updating the mockgcp for all APIs, it'd be appreciated if you could help out there! Thanks!

maqiuyujoyce

Looks like supporting poolerCount is in your last commit. Do you mind rerunning the test case locally against real and mock to ensure now the logs have no unexpected diffs?

maqiuyujoyce · 2025-12-11T01:28:09Z

mockgcp/generated/mockgcp/cloud/alloydb/v1beta/resources.pb.go

@@ -172,6 +172,8 @@ const (
 	DatabaseVersion_POSTGRES_15 DatabaseVersion = 3
 	// The database version is Postgres 16.
 	DatabaseVersion_POSTGRES_16 DatabaseVersion = 4
+	// The database version is Postgres 17.


Yes, we can follow up separately. Thanks for working on this!

maqiuyujoyce · 2025-12-11T01:33:13Z

mockgcp/generated/mockgcp/cloud/alloydb/v1beta/resources.pb.go

@@ -5316,6 +5307,8 @@ type Instance_ConnectionPoolConfig struct {
 	Enabled bool `protobuf:"varint,12,opt,name=enabled,proto3" json:"enabled,omitempty"`
 	// Optional. Connection Pool flags, as a list of "key": "value" pairs.
 	Flags map[string]string `protobuf:"bytes,13,rep,name=flags,proto3" json:"flags,omitempty" protobuf_key:"bytes,1,opt,name=key,proto3" protobuf_val:"bytes,2,opt,name=value,proto3"`
+	// Output only. The number of running poolers per instance.


Thank you, Nancy!

After looking at the issue description, I suspect it is the problem of the tool mentioned in the doc, which hasn't been updated for a while (and I actually never used it). We probably want to update our doc.

^ This is outside of the scope of this PR though.

google-oss-prow bot added the do-not-merge/work-in-progress label Nov 11, 2025

nancynh marked this pull request as ready for review November 12, 2025 00:34

google-oss-prow bot removed the do-not-merge/work-in-progress label Nov 12, 2025

maqiuyujoyce reviewed Nov 15, 2025

View reviewed changes

maqiuyujoyce self-assigned this Nov 15, 2025

nancynh force-pushed the mcp-kcc branch 5 times, most recently from 1f4d983 to 927c1e1 Compare November 19, 2025 21:42

nancynh force-pushed the mcp-kcc branch from 523b13f to db03fa9 Compare November 21, 2025 22:49

nancynh requested a review from maqiuyujoyce November 21, 2025 23:36

maqiuyujoyce mentioned this pull request Nov 27, 2025

AlloyDBInstanceSpec in krm will have QueryInsightsInstanceConfig and … #5633

Open

1 task

nancynh added 14 commits December 4, 2025 01:04

feat: add MCP support for AlloyDB

0d81fd9

update mockgcp for AlloyDB only

249f4a7

log: real gcp

1fdbac8

log: mock gcp

50e168b

chore: add connection_pool_config field to AlloyDBInstance fuzzer

77d5798

fix spacing

4bce202

mark activation_policy as unimplemented in fuzzer

735f3a3

add connection_pool_config.pooler_count field to the instance type

ac8d853

mark allocated_ip_range_override as unimplemented in the fuzzer

2244f6e

mark network_config.network as unimplemented in the fuzzer

61b6b47

fix excluding allocated ip range overrides from the fuzzer

8d46ebf

exclude unimplemented cluster fields from the fuzzer

2fbce83

add connectionPoolConfig.poolerCount to output only spec field list, …

6653b68

…and under spec for the AlloyDB instance fuzzer

log: real gcp

54e07a8

nancynh added 2 commits December 4, 2025 01:04

fix mocked pooler count to only be set when connectionPoolConfig.enab…

265ecdf

…led is true

log: mock gcp

a567082

nancynh force-pushed the mcp-kcc branch from 304623f to a567082 Compare December 4, 2025 01:04

Merge branch 'GoogleCloudPlatform:master' into mcp-kcc

c8fee53

nancynh mentioned this pull request Dec 4, 2025

Generated AlloyDB types excluded connection_pool_config.pooler_count field #5753

Open

3 tasks

use correct method for referencing output only fields for connectionP…

d7ea0f8

…oolConfig

nancynh force-pushed the mcp-kcc branch from 032e380 to d7ea0f8 Compare December 5, 2025 00:24

maqiuyujoyce reviewed Dec 11, 2025

View reviewed changes

Add Managed Connection Pooling support for an AlloyDB Instance #5611

Are you sure you want to change the base?

Add Managed Connection Pooling support for an AlloyDB Instance #5611

Uh oh!

Conversation

nancynh commented Nov 11, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

BRIEF Change description

WHY do we need this change?

Special notes for your reviewer:

Does this PR add something which needs to be 'release noted'?

Additional documentation e.g., references, usage docs, etc.:

Intended Milestone

Tests you have done

Uh oh!

nancynh commented Nov 12, 2025

Uh oh!

maqiuyujoyce left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

nancynh Dec 8, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

maqiuyujoyce Dec 11, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

google-oss-prow bot commented Nov 18, 2025

Uh oh!

nancynh commented Nov 21, 2025

Uh oh!

nancynh commented Dec 8, 2025

Uh oh!

maqiuyujoyce left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

maqiuyujoyce Dec 11, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

nancynh commented Nov 11, 2025 •

edited

Loading

nancynh Dec 8, 2025 •

edited

Loading

maqiuyujoyce Dec 11, 2025 •

edited

Loading

maqiuyujoyce Dec 11, 2025 •

edited

Loading