fix(rds): run MySQL and MariaDB with their parameter group values by drakeo338 · Pull Request #5981 · floci-io/floci

drakeo338 · 2026-10-19T05:10:14Z

Summary

Fixes #5946. Parameter group values never reached the database. Every container start (create, reboot, stop/start, Floci restart) now passes the group's overrides to MySQL or MariaDB as --loose-<name>=<value>: the DB cluster parameter group for a cluster, the DB parameter group for a standalone instance. loose makes the server skip an Aurora-only parameter with a warning instead of failing. Formula values and the listen/data-dir parameters are left out, and Floci's authentication options go last. DescribeDBClusters now reports DBClusterParameterGroup. PostgreSQL is not covered yet; the docs say so.

Type of change

Bug fix (fix:)
New feature (feat:)
Breaking change (feat!: or fix!:)
Docs / chore

AWS Compatibility

The engine reported 151 0 10 for max_connections, slow_query_log, long_query_time after the group set 500 1 1. RdsParameterGroupEngineIntegrationTest (Docker, JDBC) now sees 500 1 1, and a change after RebootDBCluster.

Checklist

./mvnw test passes locally
New or updated integration test added
Commit messages follow Conventional Commits
I have no more than 2 open, non-draft pull requests and no more than 4 open pull requests in total (drafts included) in this repository (maintainers and dependency bots are exempt).

qodo-code-review · 2026-10-19T05:10:58Z

PR Summary by Qodo

Apply RDS parameter groups when MySQL and MariaDB containers start

🐞 Bug fix 🧪 Tests 📝 Documentation 🕐 40+ Minutes

AI Description

• Apply cluster or instance parameter-group overrides whenever MySQL and MariaDB containers start.
• Preserve Floci-controlled settings and report the cluster parameter group in DB cluster responses.
• Add engine-level tests and document restart behavior and unsupported engines.

Diagram

graph TD
  H["RDS query handler"] --> S["RDS service"] --> G[("Parameter groups")] --> M["Container manager"] --> D["MySQL MariaDB"]
  S --> M
  H --> X["Cluster XML response"]

High-Level Assessment

Passing overrides as startup options fits the existing container lifecycle and supports parameters that cannot be changed at runtime. Applying changes through SQL or generated option files would add lifecycle complexity without improving this fix.

Files changed (7) +585 / -106

Bug fix (3) +165 / -14

RdsQueryHandler.javaReport DBClusterParameterGroup in cluster XML +22/-2

Report DBClusterParameterGroup in cluster XML
• Serializes the assigned cluster parameter group, or the engine-family default when no group was assigned.
src/main/java/io/github/hectorvent/floci/services/rds/RdsQueryHandler.java

RdsService.javaSupply parameter-group overrides on every container start +78/-9

Supply parameter-group overrides on every container start
• Looks up instance or cluster overrides within the resource's account and region. Passes them to container starts during creation, reboot, start, retry, and persisted-runtime restoration.
src/main/java/io/github/hectorvent/floci/services/rds/RdsService.java

RdsContainerManager.javaConvert overrides into safe database startup options +65/-3

Convert overrides into safe database startup options
• Adds MySQL and MariaDB overrides as ordered '--loose-' options before Floci's authentication options. Excludes formulas, invalid names, and settings controlling the listener or data directory; leaves other engines unchanged.
src/main/java/io/github/hectorvent/floci/services/rds/container/RdsContainerManager.java

Tests (3) +407 / -92

RdsParameterGroupEngineIntegrationTest.javaVerify parameter values in running MySQL engines +202/-0

Verify parameter values in running MySQL engines
• Uses Docker and JDBC to check cluster and standalone instance overrides against engine variables. Also checks cluster-group reporting and a changed value after reboot.
src/test/java/io/github/hectorvent/floci/services/rds/RdsParameterGroupEngineIntegrationTest.java

RdsServiceTest.javaVerify override propagation through RDS service starts +129/-84

Verify override propagation through RDS service starts
• Adds assertions for cluster creation and reboot, standalone instance creation, and starts without a group. Adjusts existing container-start mocks for the new parameter argument.
src/test/java/io/github/hectorvent/floci/services/rds/RdsServiceTest.java

RdsContainerManagerTest.javaTest startup-option filtering and container command construction +76/-8

Test startup-option filtering and container command construction
• Checks option ordering, formula and controlled-setting exclusions, MySQL and MariaDB behavior, and non-MySQL isolation. Updates existing start calls for the new argument.
src/test/java/io/github/hectorvent/floci/services/rds/container/RdsContainerManagerTest.java

Documentation (1) +13 / -0

rds.mdDocument parameter-group runtime behavior +13/-0

Document parameter-group runtime behavior

• Explains which groups configure MySQL and MariaDB, when changed values take effect, which values are omitted, and that PostgreSQL and SQL Server remain unsupported.

docs/services/rds.md

qodo-code-review · 2026-10-19T05:14:17Z

Code Review by Qodo

🐞 Bugs (0) 📘 Rule violations (0) 📎 Requirement gaps (0) 🔗 Cross-repo conflicts (0) 📜 Skill insights (0)

1. ~~A group setting can disconnect databases~~ ✓ Resolved

Description

engineParameterOptions accepts skip-grant-tables=ON as a server option even though MySQL also
disables network connections when that option is enabled. A user can store that setting through the
parameter-group API; on the next start, the database no longer accepts the TCP connections used by
Floci's proxy.

Code

src/main/java/io/github/hectorvent/floci/services/rds/container/RdsContainerManager.java[R1240-1244]

+                    || FLOCI_CONTROLLED_MYSQL_PARAMETERS.contains(
+                            name.replace('-', '_').toLowerCase(Locale.ROOT))) {
+                continue;
+            }
+            options.add("--loose-" + name + "=" + value);

Relevance

●●● Strong
Allowing a stored option to disable required networking is a direct safety gap; security-setting
validation fixes are accepted.
PR-#5304
PR-#5945

ⓘ Recommendations generated based on similar findings in past PRs

Evidence
The API accepts a supplied parameter name and value without checking whether they are safe server
options. The new filter does not exclude skip-grant-tables, and the resulting command is supplied
to Docker; MySQL documents that this option automatically enables skip_networking, while the proxy
starts against the container's network host and port.
src/main/java/io/github/hectorvent/floci/services/rds/RdsQueryHandler.java[1508-1519]
src/main/java/io/github/hectorvent/floci/services/rds/container/RdsContainerManager.java[65-68]
src/main/java/io/github/hectorvent/floci/services/rds/container/RdsContainerManager.java[1228-1245]
src/main/java/io/github/hectorvent/floci/services/rds/container/RdsContainerManager.java[310-337]
src/main/java/io/github/hectorvent/floci/services/rds/RdsService.java[3516-3539]
🌐 MySQL states that starting with --skip-grant-tables also enables skip_networking, disabling remote connections.

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
A parameter-group override can indirectly disable the network connection required by the RDS proxy.
## Fix Focus Areas
- src/main/java/io/github/hectorvent/floci/services/rds/container/RdsContainerManager.java[65-68]
- src/main/java/io/github/hectorvent/floci/services/rds/container/RdsContainerManager.java[1228-1245]
## Recommended Fix
Exclude `skip-grant-tables` and other server options that implicitly disable networking or authentication, preferably by validating overrides against supported RDS parameters. Add a test that the dangerous option is not passed to the container.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

2. ~~Unexpected Docker errors skip the tests~~ ⊘ Outdated

Description

checkDocker() catches every Exception from the Docker ping and converts it into an assumption
failure instead of preserving unexpected failures. Any programming, configuration, or Docker-client
error during the health check is therefore reported as an unavailable Docker daemon and the
integration test is silently skipped.

Code

src/test/java/io/github/hectorvent/floci/services/rds/RdsParameterGroupEngineIntegrationTest.java[R43-45]
+        } catch (Exception e) {
+            assumeTrue(false, "Docker is not available: " + e.getMessage());
+        }

Relevance

●●● Strong
An exact recent precedent accepted changing broad Docker-ping exception handling in an integration
test.
PR-#5945

ⓘ Recommendations generated based on similar findings in past PRs

Evidence
The changed test catches broad Exception and unconditionally calls assumeTrue(false), so every
failure from the Docker ping becomes a skipped test. This violates the requirement to suppress only
known benign errors.
Rule 3865600: Do not swallow exceptions broadly; suppress only known benign errors
src/test/java/io/github/hectorvent/floci/services/rds/RdsParameterGroupEngineIntegrationTest.java[40-45]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
The Docker availability check catches every exception and skips the test, masking unexpected test or client failures as Docker unavailability.

## Fix Focus Areas
- src/test/java/io/github/hectorvent/floci/services/rds/RdsParameterGroupEngineIntegrationTest.java[40-45]

## Recommended Fix
Catch only the specific Docker-client or connection exceptions that indicate the daemon is unavailable, and allow unexpected runtime or configuration exceptions to propagate so the test fails visibly.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

3. ~~Group switches leave old values in use~~ ✓ Resolved

Description

instanceEngineParameters and clusterEngineParameters read the persisted group name, but the
corresponding modify handlers never accept or persist a replacement group name. When a client
switches an existing instance or cluster to another parameter group through ModifyDBInstance or
ModifyDBCluster, a subsequent reboot still starts the engine with the originally associated
group's values.

Code

src/main/java/io/github/hectorvent/floci/services/rds/RdsService.java[R8846-8848]

+    private Map<String, String> clusterEngineParameters(DbCluster cluster) {
+        return clusterEngineParameters(accountIdFromArn(cluster.getDbClusterArn()),
+                regionFromArn(cluster.getDbClusterArn()), cluster.getParameterGroupName());

Relevance

●●● Strong
Modify operations must persist replacement groups for later restarts; repository accepts fixes for
inconsistent persisted state.
PR-#4332
PR-#5787

ⓘ Recommendations generated based on similar findings in past PRs

Evidence
Neither modify handler passes a parameter-group name into its service modification method. The new
startup helpers use only the model's stored name, so the requested replacement cannot reach the
engine on any later start.
src/main/java/io/github/hectorvent/floci/services/rds/RdsQueryHandler.java[732-770]
src/main/java/io/github/hectorvent/floci/services/rds/RdsQueryHandler.java[1396-1430]
src/main/java/io/github/hectorvent/floci/services/rds/RdsService.java[8815-8819]
src/main/java/io/github/hectorvent/floci/services/rds/RdsService.java[8845-8849]
src/main/java/io/github/hectorvent/floci/services/rds/RdsService.java[3762-3768]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
Startup now uses the resource's persisted group, but modification requests cannot change that association.
## Fix Focus Areas
- src/main/java/io/github/hectorvent/floci/services/rds/RdsQueryHandler.java[732-770]
- src/main/java/io/github/hectorvent/floci/services/rds/RdsQueryHandler.java[1396-1430]
- src/main/java/io/github/hectorvent/floci/services/rds/RdsService.java[8815-8819]
- src/main/java/io/github/hectorvent/floci/services/rds/RdsService.java[8845-8849]
## Recommended Fix
Parse the group-name fields in both modify handlers, validate their engine family and scope, and persist the association before subsequent backend starts. Test switching to a second group and rebooting.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

4. ~~One rejected parameter value stops the database~~ ✓ Resolved

Description

engineParameterOptions passes every non-formula value as --loose-<name>=<value>, but loose
only suppresses unknown option *names*. A known option with a value the server rejects still stops
mysqld: for example the Aurora-only binlog_format=OFF, a path such as log_error under an RDS
directory, or lower_case_table_names that differs from the already-initialised data dictionary.
When that happens on create, reboot, stop/start or Floci restart, the container never comes up and
the cluster or instance goes to FAILED. This includes deployments that already had such a group
stored before this change.

Code

src/main/java/io/github/hectorvent/floci/services/rds/container/RdsContainerManager.java[R1238-1244]

+            if (name == null || value == null || !MYSQL_PARAMETER_NAME.matcher(name).matches()
+                    || value.contains("{")
+                    || FLOCI_CONTROLLED_MYSQL_PARAMETERS.contains(
+                            name.replace('-', '_').toLowerCase(Locale.ROOT))) {
+                continue;
+            }
+            options.add("--loose-" + name + "=" + value);

Relevance

●●● Strong
Startup reliability findings involving unsafe or rejected database settings align with accepted
container robustness fixes.
PR-#5741
PR-#5304

ⓘ Recommendations generated based on similar findings in past PRs

Evidence
The filter only drops formula values, invalid names and the five listen/data-dir parameters. Every
other value goes into the container command unchanged, and every start path (create, reboot, ensure,
restore) now passes the stored group parameters. The docs also say that only *unknown* parameters
are skipped, so they do not warn users that a bad value stops the database.
src/main/java/io/github/hectorvent/floci/services/rds/container/RdsContainerManager.java[1228-1247]
src/main/java/io/github/hectorvent/floci/services/rds/RdsService.java[8094-8098]
docs/services/rds.md[517-526]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
Parameter group values are passed to mysqld/mariadbd as `--loose-name=value`. `loose` does not cover invalid values or settings that conflict with the data directory (for example `lower_case_table_names`), so a single bad value stops the server and the RDS resource goes to FAILED, including on the first Floci restart after an upgrade.

## Fix Focus Areas
- src/main/java/io/github/hectorvent/floci/services/rds/container/RdsContainerManager.java[1228-1247]
- docs/services/rds.md[517-526]

## Recommended Fix
Add `lower_case_table_names` (and file/path parameters such as `log_error`, `pid_file`, `init_file`, `secure_file_priv`) to the Floci-controlled or skipped set. Also consider retrying the start without the group options, logging which ones were dropped, if the container exits during startup. Update the docs to say that a value the server rejects stops the database.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

5. ~~Some clusters report nonexistent groups~~ ✓ Resolved

Description

dbClusterParameterGroupName constructs a default. group name from every cluster engine's family
without checking whether the service provides that group. For a MariaDB or SQL Server cluster
created without an explicit group, DescribeDBClusters consequently names a group that
DescribeDBClusterParameterGroups cannot list and getDbClusterParameterGroup cannot retrieve.

Code

src/main/java/io/github/hectorvent/floci/services/rds/RdsQueryHandler.java[R3069-3072]

+        try {
+            return "default." + service.expectedClusterParameterGroupFamily(engine, cluster.getEngineVersion());
+        } catch (AwsException e) {
+            return null;

Relevance

●●● Strong
Engine-specific default-group reporting should avoid synthesizing nonexistent groups; recent engine
restriction fixes were accepted.
PR-#5736
PR-#5206

ⓘ Recommendations generated based on similar findings in past PRs

Evidence
Cluster creation supports MariaDB and SQL Server, and the family calculator can produce names for
both. Neither family occurs in the managed cluster-group list, while retrieval resolves only stored
or managed groups; the newly constructed fallback therefore points to no group.
src/main/java/io/github/hectorvent/floci/services/rds/RdsQueryHandler.java[3064-3072]
src/main/java/io/github/hectorvent/floci/services/rds/RdsService.java[115-140]
src/main/java/io/github/hectorvent/floci/services/rds/RdsService.java[7185-7199]
src/main/java/io/github/hectorvent/floci/services/rds/RdsService.java[7241-7270]
src/main/java/io/github/hectorvent/floci/services/rds/RdsService.java[6681-6694]
src/main/java/io/github/hectorvent/floci/services/rds/RdsService.java[6697-6714]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
DescribeDBClusters invents default group names for engines whose default groups are not available through the service.
## Fix Focus Areas
- src/main/java/io/github/hectorvent/floci/services/rds/RdsQueryHandler.java[3064-3073]
- src/main/java/io/github/hectorvent/floci/services/rds/RdsService.java[115-140]
## Recommended Fix
Only report a fallback default when it resolves to a managed group, or add the appropriate defaults to the service's managed group catalog. Test that every reported fallback can be described by name.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

6. ~~A group's time zone setting never takes effect~~ ✓ Resolved

Description

engineParameterOptions uses the RDS parameter name as the command-line option, but some MySQL
system variables do not have a command-line option of the same name. For example, time_zone only
exists as --default-time-zone, so --loose-time_zone=UTC is skipped with a warning. Even though
the ticket names time_zone as a common need, a group that sets it leaves the engine at SYSTEM,
which contradicts the docs' statement that "a value set in the group is the engine's value".

Code

src/main/java/io/github/hectorvent/floci/services/rds/container/RdsContainerManager.java[1244]
+            options.add("--loose-" + name + "=" + value);

Relevance

●●● Strong
Directly translating an RDS parameter into an invalid server option contradicts the feature contract
and should be corrected.
PR-#5741

ⓘ Recommendations generated based on similar findings in past PRs

Evidence
The option is built directly from the parameter name. MySQL's command-line and option-file form of
time_zone is default_time_zone, and the server reports time_zone as an unknown variable, so
the loose- form is silently skipped.
src/main/java/io/github/hectorvent/floci/services/rds/container/RdsContainerManager.java[1235-1245]
docs/services/rds.md[517-523]
🌐 Using time_zone in an option file gives "unknown variable 'time_zone=...'"; the option-file format is default_time_zone.

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
Some RDS parameter names do not have a mysqld command-line option of the same name (for example `time_zone` -> `--default-time-zone`), so those values are silently ignored.

## Fix Focus Areas
- src/main/java/io/github/hectorvent/floci/services/rds/container/RdsContainerManager.java[1235-1245]
- docs/services/rds.md[517-526]

## Recommended Fix
Add a small rename map (for example `time_zone` -> `default-time-zone`) before building the option. Only offset values like `+00:00` or `UTC` are safe without tz tables, so either check the value or document the limitation in rds.md.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

Context sources

✅ Tickets:

🎫 [BUG] RDS: DB cluster parameter group values are never applied to the MySQL engine, and DescribeDBClusters does not report the group

✅ Compliance rules (platform): 8 rules

✅ Web pages:

🌐 MySQL :: MySQL 8.0 Reference Manual :: 8.2.17 Pluggable Authentication

🌐 MySQL :: MySQL 8.0 Reference Manual :: 8.2.22 Troubleshooting Problems Connecting to MySQL

🌐 MySQL Bugs: #90554: Undesired change for Windows users in 8.0

+7 more

✅ Cross-repo context — repo relationships

Explored: repo: floci-io/floci-ui (sha: bb3c8e0e) — View relationship

Review mode: Auto: 🧠 Deep: Many lifecycle paths and parameter lookups create independent defect opportunities.

Tip of the day

💡 Did you know, you can choose which labels appear on a finding, and whether they show icons or text

More tips ↗ | Customize Qodo ↗ | Qodo docs ↗

pgermosen

Thanks for picking this up. It's been a real gap.

One thing worries me before the rest. Plenty of hardened setups turn on require_secure_transport, and Floci's proxy talks to MySQL without TLS, so once that value reaches the server nobody can connect, TLS clients included. skip_grant_tables ends the same way because it switches networking off. I'd skip both, like port and socket.

Aurora is the other gap. max_connections normally sits in the instance's own parameter group, which overrides the cluster one, but a member here never passes its group to the database. The writer's group should be layered on top.

#5946 also asks for switching groups with ModifyDBCluster and ModifyDBInstance, and those still ignore the new name.

drakeo338 · 2026-10-20T05:15:10Z

Pushed 8d1d6e7: skip require_secure_transport and skip_grant_tables like port and socket; Aurora clusters layer the writer's parameter group over the cluster group; ModifyDBCluster and ModifyDBInstance apply a new parameter group from the next reboot.

pgermosen

@drakeo338 thanks, all three are in: both settings are skipped, the writer's group goes on top, and the modify calls switch groups. A few things left. On Aurora, AWS applies a changed group when the instances are rebooted (RebootDBCluster is "only for a non-Aurora Multi-AZ DB cluster"), but RebootDBInstance on an Aurora member doesn't restart the cluster's container, so the change never lands that way; the docs and the integration test use RebootDBCluster. time_zone has no command-line option in MySQL (it's --default-time-zone), so a group's time zone is skipped with a warning. A value the server rejects, such as Aurora's binlog_format=OFF, keeps mysqld from starting while the instance still reports available, where AWS reports incompatible-parameters. And the "Written with help from Claude Code" line needs to come out of the description, since attribution stays with human contributors.

(follow-up, separate PR) #5946 also expects a dynamic parameter changed with ApplyMethod=immediate to apply without a reboot, and ParameterApplyStatus stays in-sync after a change, so I'd use "Refs #5946" rather than closing it.

A DB cluster parameter group or DB parameter group was stored but never reached the engine: the container started with only the authentication option, so the database kept MySQL's built-in defaults. Every container start (create, reboot, stop and start, emulator restart) now passes the group's overrides as --loose-<name>=<value> server options: the cluster parameter group for a cluster, the DB parameter group for a standalone instance. The loose prefix makes the server skip an Aurora-only parameter with a warning. Formula values and the parameters that decide where the server listens and keeps its data are left out, and Floci's authentication options come last so the proxy keeps working. DescribeDBClusters now reports DBClusterParameterGroup. Fixes floci-io#5946

… on modify Review follow-up: - require_secure_transport and skip_grant_tables are no longer passed to the server, like port and socket: Floci's proxy talks to MySQL without TLS, and skip_grant_tables turns networking off, so either would lock every client out. - An Aurora cluster's engine now runs with its DB cluster parameter group and the writer instance's DB parameter group on top, so an instance-level value such as max_connections reaches the database. Creating the writer restarts the cluster's container when its group changes the engine's options; every later start reads the writer's group too. - ModifyDBCluster takes DBClusterParameterGroupName and ModifyDBInstance takes DBParameterGroupName. The group is checked against the engine before anything else in the request is applied, recorded at once, and used from the database's next start (a reboot), as on AWS.

…ble-parameters Review follow-up: - RebootDBInstance on an Aurora cluster's writer restarts the database the cluster's members share, which is how AWS applies a changed DB cluster parameter group or writer DB parameter group. A reader's reboot still only restarts its endpoint. - time_zone reaches the server as --default-time-zone, since MySQL has no time_zone option, and Aurora's binlog_format=OFF, which turns binary logging off and which MySQL refuses as a value, as --skip-log-bin. - A server that stops while starting with its parameter group's options leaves no container, and its instance, or every member of its cluster, reports incompatible-parameters until a reboot starts it with values the server accepts. Floci logs the server's error lines. - DescribeDBClusters names a default DB cluster parameter group only when DescribeDBClusterParameterGroups lists it. - The integration test reboots the writer rather than the cluster, checks that binary logging is off, covers a refused value, and skips through RdsRequests.assumeDocker.

pgermosen

@drakeo338 thanks for sticking with this one. A group's values now reach MySQL and MariaDB on every start, rebooting the writer applies an Aurora cluster's changes, and a value the server refuses shows up as incompatible-parameters rather than an instance that looks available but never answers.

drakeo338 requested review from hectorvent and pgermosen as code owners October 19, 2026 05:10

qodo-code-review Bot reviewed Oct 19, 2026

View reviewed changes

hectorvent added bug Something isn't working rds Amazon Relational Database Service (RDS) labels Oct 19, 2026

pgermosen reviewed Oct 19, 2026

View reviewed changes

pgermosen added the waiting-author Review posted; waiting on the PR author to respond label Oct 19, 2026

pgermosen reviewed Oct 20, 2026

View reviewed changes

drakeo338 and others added 3 commits October 23, 2026 21:00

pgermosen force-pushed the fix/rds-apply-parameter-groups branch from 8d1d6e7 to adeeab5 Compare October 24, 2026 01:02

pgermosen approved these changes Oct 24, 2026

View reviewed changes

pgermosen removed the waiting-author Review posted; waiting on the PR author to respond label Oct 24, 2026

pgermosen merged commit 16c159e into floci-io:main Oct 24, 2026
26 checks passed

Navigation Menu

Repository navigation

fix(rds): run MySQL and MariaDB with their parameter group values - #5981

fix(rds): run MySQL and MariaDB with their parameter group values#5981
pgermosen merged 3 commits into
floci-io:mainfrom
drakeo338:fix/rds-apply-parameter-groups

drakeo338 commented Oct 19, 2026 •

edited by pgermosen

Loading

Uh oh!

qodo-code-review Bot commented Oct 19, 2026

Uh oh!

qodo-code-review Bot commented Oct 19, 2026 •

edited

Loading

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

pgermosen left a comment

Uh oh!

drakeo338 commented Oct 20, 2026

Uh oh!

pgermosen left a comment

Uh oh!

pgermosen left a comment

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

drakeo338 commented Oct 19, 2026 • edited by pgermosen Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Type of change

AWS Compatibility

Checklist

Uh oh!

qodo-code-review Bot commented Oct 19, 2026

PR Summary by Qodo

Uh oh!

qodo-code-review Bot commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Code Review by Qodo

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

pgermosen left a comment

Choose a reason for hiding this comment

Uh oh!

drakeo338 commented Oct 20, 2026

Uh oh!

pgermosen left a comment

Choose a reason for hiding this comment

Uh oh!

pgermosen left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

drakeo338 commented Oct 19, 2026 •

edited by pgermosen

Loading

qodo-code-review Bot commented Oct 19, 2026 •

edited

Loading