Changelog

We've have added support for instant deployment of certain deploy requests using MySQL's ALGORITHM=INSTANT. This feature allows users to apply changes more quickly and efficiently.

We've shipped a change to service token permissions. Now, any service token with create_databases, will automatically be granted full permissions to any database that it creates.

With this change, you can fully automate creation + management without needing to grant more permissions to your service token.

We’ve added a new feature to the top navigation bar that lets you choose between displaying timestamps in your local time zone or UTC. Making it easier to communicate and share graphs with your teammates in various timezones.

Your preference will be saved and persist across browser sessions.

We've added a new column in the main Insights view: max_latency. This column shows the maximum response time of any individual query pattern in the selected time window. Maximum is often a useful metric for spotting extreme outliers that may not show up in p99 (99th percentile). To enable the Max latency column, visit your Insights dashboard, click on the columns menu to the right of the query filter search box, and check "Max latency".

You can also filter query patterns based on max latency by adding max_latency:>x where x is measured in miliseconds.

We've just shipped a couple improvements to branching. Now from the branches page, you can:

• Rename any of your branches.
• Replace your default branch with another branch (admins only).

When using "Set as default branch" you can optionally choose to have the children of the current default branch move over with it.

Permissions:

• Database admins are able to rename any branch as well as change the default.
• Database members may only rename development branches.

With these changes, you now have more control over the organization of your branches.

Insights now stores index usage information for all SELECT queries. To see index usage, click on a query from your Insights dashboard, and then click on the new Indexes tab.

This view shows a time series graph of the percent of queries that use each listed index. The vertical bar chart near the bottom shows comprehensive index usage information for the entire period.

You can also search for queries based on index usage:

• To find queries that use a given index, add index:$TABLE.$INDEX_NAME to the insights search bar
• To find queries that don't use any indexes try indexed:false

In addition to index usage, we've also added summary statistics (total time, latencies, etc) to the query details page.

Read the blog post for more information.

Backups are deleted after they expire, but automatic and manual deletions can be prevented by enabling deletion protection.

• /v1/organizations/{organization}/databases/{database}/branches/{branch}/backups/{id}

PlanetScale's latest SOC 2 Type II report covering the period of June 16, 2023 to June 15, 2024 is now available to customers.

To receive a copy of the report, please contact Support or visit our Trust Portal.

We've heard your feedback and Insights graphs now show all x-axis time measurements as an absolute time (e.g. 09:31) instead of as a relative time offset (e.g. 3h ago). Absolute time offsets make it easier to correlate Insights data with information from external systems.

We've increased the resolution for all Insights graphs. Graphs with periods from 12 to 24 hours now show data in 5 minute increments (down from 10 minutes) and periods under 12 hours show data in 1 minute increments (down from 5 minutes). Smaller time buckets result in less smoothing and make it easier to spot temporary fluctuations.

We have shipped a new webhook event: deploy_request.pending_cutover.

This event will trigger when a deploy request is ready to apply the schema and is waiting on the user to confirm.

For deploy requests that have "auto_cutover" or "Auto apply" enabled, this webhook will not be sent.

This new event is now available in your webhook settings.

We just shipped an improvement to deploy requests. Now, you can optionally set your deploy requests to delete the development branch once fully complete.

You can do this by toggling a checkbox while the deploy request is running. This new setting is also available via the CLI and API.

It is disabled by default and must be enabled for each deploy request. The branch will only be deleted if the deployment is successful and the revert period is complete.

CLI

As of pscale version 0.205.0, you can create a deploy request with pscale deploy-request create db branch --auto-delete-branch and the setting will be enabled.

API

The create deploy request endpoint now has auto_delete_branch available. Set this parameter to true to enable the setting.

Two years ago we made an upgrade to our edge infrastructure that enabled better performance and reliability for your database connections.

A small number of databases are still using connection strings from before this change. In the coming weeks we are working on deprecating the usage of these strings and getting everyone upgraded.

Does this impact my database?

We are sending emails to all account admins for databases which are still connecting via legacy strings.

You may also check the hostname for your connections tring. If it shows the ID for the database branch, then it is a legacy string.

Examples:

• Legacy host name: abcdeabc1234.us-east-1.psdb.cloud
• New infrastructure host name: aws.connect.psdb.cloud

How to update my connection string?

To update your connection string, go to the PlanetScale dashboard -> your database -> settings -> passwords.

Create a new password and update the hostname, username and password in your application.

What happens if I do not upgrade?

We will be performing multiple brownout periods starting September 3rd, 2024. During these, connections to your database will stop working.

The purpose of the brownout periods is to help alert anyone who is still unintentionally using a legacy connection string.

After several brownout periods, we will re-evaluate usage again and send final notices to anyone still using a legacy connection string before fully deprecating.

pscale CLI

If you are using an old version of the pscale CLI, you may see the following message when trying to use pscale shell or pscale connect.

Your version of pscale is too old and is using a deprecated connection method. Please update your CLI version.

If this impacts you, please update pscale to the latest version.

Need help?

We want this process to be simple and painless for everyone. If this causes any difficulties for you, please reach out to our support team.

We've shipped a small update to the "Require administrator approval for deploy requests" setting. Now, if your database only has a single administrator, they can self-approve their own deploy requests.

If there is more than one administator, then self-approval is not allowed.

We have made an update to schema recommendations. Now when closing a recommendation, you can leave an optional comment explaining why the recommendation is being closed.

Comments support markdown.

We have updated all webhook payloads to include the database and organization names.

{
  "timestamp": 1698252879,
  "event": "branch.ready",
+  "organization": "my-org",
+  "database": "example_database",
  "resource": { ... }
}

We have released a new API endpoint for retrieving your organization's audit logs.

• /v1/organizations/:name/audit-log

A new service token scope read_audit_logs has also been added. You may add this to your service token by visiting your service token settings.

We are removing support for versions of the PlanetScale CLI older than v0.182.0 (February 2, 2024). On July 1st, they will no longer be able to use the following commands:

• pscale connect
• pscale shell
• pscale database dump
• pscale database restore-dump

Please update your CLI to version v0.182.0 or later to continue using these commands.

Branch cluster size has been added to the PlanetScale API. Now you can programmatically change your cluster size via a service token. To use the endpoints, you will need the write_database permission.

• /v1/organizations/{organization}/databases/{database}/branches/{name}/cluster

We’ve just shipped a change to the layout of the PlanetScale app. The top-level navigation links are now along the left-hand side.

Why make this change? We needed to make room for new features. We were starting to be limited by the space available along the top navigation bar. The new side navigation allows for more space and keeps everything easily discoverable.

We are now publishing the rate of errors encountered by each database branch's VTGate to Datadog. You can graph these metrics by using the planetscale.vtgate.errors metric name within your Datadog widget and notebooks.

We've shipped an improvement to make navigating to your most frequently used databases easier. Starting today, you can pin the databases you use most frequently. To do this, open the database switcher at the top of the page, you'll see a new pin icon next to each database.

The organization switcher has been improved as well. It will show all of your pinned databases across all of your organizations.

Invoices have been added to the PlanetScale API. Now you can programmatically access your organizations billing data via OAuth or a service token. To use the endpoints, you will need the read_invoices permission.

• /v1/organizations/{organization}/invoices
• /v1/organizations/{organization}/invoices/{id}
• /v1/organizations/{organization}/invoices/{id}/line-items

Previously only available to Single-tenant and Managed customers, PlanetScale now provides private connection endpoints to GCP regions via GCP Private Service Connect for customers on Scaler Pro.

Previously only available to Single-tenant and Managed customers, PlanetScale now provides private connection endpoints to AWS regions via AWS PrivateLink for customers on Scaler Pro.

This is helpful in cases where your compliance mandates that your connections do not route through the public Internet, or for customers looking to save costs on their NAT Gateways.

Insights now supports filtering query patterns by 50th and 99th percentile query execution time. To find all query patterns that have a 50th percentile execution time above 250 milliseconds, for example, enter p50:>250 in the insights search box.

We now detect usage of legacy character sets and collations in your database and provide upgrade recommendations. Modern character sets and collations are faster, support a wider range of characters including emoji, and have improved sorting for multi-byte characters.

Today, we're introducing the ability to create global replica credentials for your PlanetScale database.

With replica credentials, it makes it easier to query your database's replicas and read-only regions by creating a password dedicated for replica use. These new credentials will be routed to the nearest read-only region with the lowest latency as new regions are added or removed without any code changes or even needing to reconnect.

Webhooks have graduated from beta and are now generally available. Since the beginning of beta we've added an additional event which will trigger whenever a new Insights anomaly event is detected.

Our Hobby plan has now been retired. All remaining databases on the Hobby plan have been slept. To access your data, you must either upgrade to Scaler Pro or wake your database from your dashboard. Waking your database will give you 24 hours to dump your data. Please view our Hobby plan deprecation documentation for more information.