View response times (percentiles/averages), memory consumption, request numbers and error rates at the application and service level. Data is collected at one minute granularity.

See performance metrics aggregated for SQL, HTTP, MongoDB, Redis and other downstream databases and services.

View a list of the slowest or highest impact endpoints for each service including average and problem response times, memory consumption and error rates.

View average response time, percentiles in total and for individual downstream layers, the full distribution of requests, number of requests and error rate down to one minute granularity for each individual endpoint.

See performance metrics aggregated for SQL, HTTP, MongoDB, Redis and other downstream databases and services on the endpoint level.

With service monitoring you can separate frontend from admin-backend, API, worker jobs, or microservices. Each endpoint is assigned a service, and aggregated performance data is calculated for all endpoints that report for each service.

A Tideways production environment is meant to be connected to the production setup of a project. It stores data with the regular retention of the license and keeps several traces per minute from more than one host.

Collect metrics and traces for non-production environments such as staging systems, user acceptence testing environments, developer machines or pull request review apps to get insights from the full application lifecycle. All these non-production environments have only 1 day metrics and trace retention.

Track the performance of your cron- and background worker jobs as well as profile them using trace sampling.

With application centric Slow SQL Monitoring Tideways provides similar functionality to MySQL and PostgreSQL Slow Query logs, with the additional benefit of getting code context and stacktraces where the slow SQL were executed in your project, de-duplication of repeated slow SQL queries and a workflow to close or ignore slow SQL queries that you don't need to know about. Slow SQL Monitoring is available at 1 second resolution for request-based licenses and 250ms for basic licenses and 100ms for Standard and Pro licenses.

Number of endpoints that cab be monitored in a project. Endpoints exceeding this limit are grouped together.

Track deployments with detailed reports that compare before and after performance.

Get notified about alerts, events and incidents in your project with messages send through one of our many integrations, for example Slack, Microsoft Teams, OpsGenie, PagerDuty, Webhooks and E-Mail.

Configure response time and error-rate alerts for individual endpoints.

In addition to response time/latency tracing, you can also track memory usage of your timespans, or at the function level with Tideways Memory Profiler.

Track wall-time, memory and number of invocations for every function call in your PHP program. Drill down the callstack using tabular or graphical representations and find the functions that take the most time to execute. Cross-reference function calls with SQL queries, HTTP calls, templates, events and listeners from frameworks.

Tideways regularly collects timeline traces from real production requests at very low overhead. Traces contain detailed information about SQL queries, HTTP calls, NoSQL databases, PHP frameworks, Magento, Symfony, Shopware, Wordpress and many more.

Tideways already includes broad support for many well- and lesser known PHP libraries, frameworks and extensions by default. No customization is needed to Tideways to get in depth insights into PHP projects directly after installation.

Tideways selects how many traces will be stored in the backend based on the selected license. This is the largest trade-off we have to make in terms of insights for costs.

The "slow traces" sampling strategy collects traces that are slower than the 95% percentile response time of the application. Traces that are faster can only be collected when using the developer trace triggering functionality.

The "least recently traced" sampling strategy prioritizes traces from endpoints where no trace has been collected for the longest amount of time and can then keeps both slow and average duration traces. This provides a good spread of relevant traces across an application.

Modify the trace sampling strategy to prioritize traces for selected endpoints through the UI, the API or with the Tideways CLI.

Trigger the more detailed Callgraph Profiler from the UI by configuring dynamic tracepoints. When scripts of the application hit all the conditions of a tracepoint at runtime then the callgraph profiling is automatically triggered and activated.

Depending on your plan, Tideways stores traces for a short or long time frame to search and analyze. The retention is given as an upper limit, as we keep some faster traces around for a shorter amount of time in return for being able to store a larger amount of different traces for short-term analysis.

Tideways can collect stacktraces that lead up to a span in the Timeline Profiler. This helps identifying what the source is for executed SQL Queries, HTTP requests and other timespans shown in the Timeline Profiler.

Tideways can collect stacktraces that lead up to a span in the Timeline Profiler. This helps identifying where SQL Queries and other calls are coming from.

Add custom information to every trace for example a user or customer id and any context information that is necessary to better understand traces.

Generate timeline and callgraphs across multiple connected PHP requests to analyze the performance of distributed systems.

Exceptions and Fatal Errors in your PHP programs are collected including a stacktrace, context information, optionally code snippets and are deduplicated to alert you of everything that is going wrong in your project.

How many days is error/exception data stored in Tideways?

Tideways can keep a history of stacktraces and context information for errors, so that you can look at different versions of the same error for better understanding and fixing.

Directly see a snapshot of the code responsible for an exception from Tideways. The Tideways daemon annotates all lines in a stacktrace with 5 lines of code around in a frame.

Get notified about runtime deprecation errors that your code is triggering using PHPs internal deprecation message levels E_DEPREATED and E_USER_DEPRECATED. This works even when your application disables deprecation tracking on the php.ini level, so that you can safely gather data about your applications next PHP or framework update.

The weekly report gives you a high-level view on the performance of your projects including a comparison with previous historical data.

Track application, service and endpoint performance over long periods of time and compare them on a daily, weekly or monthly basis. If you subscribe to the Standard or Pro plan for your project, then historical information is tracked for primary in-depth endpoints as well.

Is endpoint performance data aggregated on a daily basis part of the historical data?

Store a number of traces preferably with callgraph data from the transactions with the highest impact every day.

The history data includes a list of Exceptions and their number of occurrences.

Other tools are billing per host/server, which we do not think makes sense and can get out of hand quickly. So we invested significant engineering effort to come up with a better model that is almost independent from host limits. When you exceed the soft limit for hosts we will start throttling your data intake. However you can easily circumvent the throttling by installing an additional proxy package in your infrastructure that pre-aggregates data before sending it to us. Using the proxy and connecting an an unlimited amount of servers to Tideways has no extra costs.

Export response times to StatsD compatible services including DogstatsD.

Access near-realtime and historic data directly from the REST API to integrate them into your other tools and processes. Different rate limits apply based on the projects plan.

How many requests per hour against the API before hitting the rate limit?

Reach our development and support team via the e-mail address [email protected]. Our active support time is on weekdays between 8:00 and 18:00 in Europe/Berlin timezone.

A 90 minute onboarding and training workshop via Zoom with your team. You pick the content you want to hear about and we will work and optimize directly on your own Tideways projects. Workshops are in German or English language. A recording for future view can be made available.