Tideways 2025.4 Release
In our fourth Release of 2025, we included PHP 8.5 support on the day of its release and Heartbeats to monitor your application’s pulse more closely than ever. If something is offbeat, you’ll receive timely alerts.
We improved alerting by introducing fine-granular transaction level response time and made automatic tracepoint triggers more powerful.
Summary
- PHP 8.5 Support
- Heartbeat Monitoring per Transaction
- Improved Alerting on Transaction Response Times
- Improvement Transaction List Filters
- More nodes in Callgraphs
- Automatic Tracepoints Trigger
- More Stacktraces on Spans
- PDO Transactions
- Incoming Network for Predis
- Magento Sections
- Minor PHP Extension Improvements
PHP 8.5 Support
PHP 8.5 has been released last week and Tideways – as always – supports it from the start with the PHP Extension Release 5.30.0.
Heartbeat Monitoring per Transaction
Some transactions (think cron jobs, payment completion, webhooks) are so critical for web applications that it is a sign of failure or at least a problem when no data was processed for a few minutes or hours.
Tideways can now send notifications on missing transaction check-ins at a granularity of 1 minute and more. For any given transaction, you can configure an individual check-in interval in the transaction settings.

Furthermore, we have renamed the “Missing Data” notifications on the service level to “Heartbeat Monitoring” because this name is clearer and well established.
Check our documentation for more detail.
Improved Alerting on Transaction Response Times
Alerting for response time thresholds on a per-transaction basis was previously exclusive to only bigger plans. We have completely rebuilt how it works to allow a much wider audience to benefit from this feature.
You can now configure the notification under transaction settings, just like all the other transaction-level notifications.

You will notice an added line in the performance chart for the transactions on which you enabled this feature.

When the response is slower than expected for 5% of all requests, then an incident is created and notifications may be sent:

Improvement Transaction List Filters
Because of these new transaction notification options for heartbeats and response times, we worked on a few small improvements for the Transaction settings list:
- You can now see at a glance which notifications are set per transaction.
- There are filters available to only select transactions with selected notifications configured.

More nodes in Callgraphs
Callgraphs can become quite complex quickly. After all, every function call is tracked. With modern frameworks or ecommerce solutions, this can be hundreds or thousands of different functions being recorded.
Visualizing this data can be a challenge. Which is why we limit the number of visible nodes by a simple algorithm: take the three most costly function calls and build the graph upwards from there. The full list of function calls has always been available using the table on the left anyway.

But since it is helpful in some cases to see a wider range of nodes, we added an option to change this start-count using a dropdown. This might be useful, e.g., with long-running scripts, where there might be more than a handful of important calls.

Automatic Tracepoints Trigger
After introducing automatic tracepoints in the last release, there is now another strategy to create a tracepoint automatically: It picks one of the seven most high-impact transactions that are rotated every week.
This will provide you with callgraph insights into different parts of the application automatically, and in combination with the trace history this data is available over a long period of time as well.
This strategy is available for Business and Enterprise plans and is used automatically for triggers created for new services. Alternatively you can select the strategy from the settings called “Rotate between highest impact transactions”.

More Stacktraces on Spans
We’ve changed the default INI settings that control when stack traces are added to spans in a trace to 1 ms (from previously 3 ms) and increased the stack depth to 10 (previously 5). Overall, this should make stack traces more valuable for pinpointing performance problems.


PDO Transactions
Long-running database transactions can have a significant impact on database performance and should thus be active as long as necessary but as short as possible. Tideways already included individual spans for BEGIN and COMMIT / ROLLBACK queries, indicating the start and end of a transaction and the COMMIT performance.
Now Tideways will also create a “meta span” spanning the entire transaction, making it easier to see at a glance which queries are part of the transaction and for how long the transaction was active. Since rollbacks can be particularly expensive and generally indicate issues, the spans for rolled-back transactions will also be marked as “errored”.

Incoming Network for Predis
When using the predis/predis library, spans will now contain a Network (incoming) annotation, just like HTTP calls or MySQL queries.

Magento Sections

To improve the efficiency of the page cache, Magento retrieves dynamic user-specific components of the resulting page in bulk from a special “Section Loader” endpoint. Traces for this endpoint now contain spans for each component loaded, helping you find out which components are particularly expensive to calculate.
Minor PHP Extension Improvements
This release, as always, contains some minor improvements to various features. Amongst those is extending support for the automated worker instrumentation to Yii 2 Queues, support for transaction name detection with Slim 4 and extended detection of email addresses in error messages and URLs in Tideways Daemon to prevent them from leaving your infrastructure.
Check our detailed changelog for more detail.
Time to get started with Profiling! With our free trial.