For the CTO and the head of product
We log every call to your API, with who made it, from which screen and how long it took, and you see it in your own dashboard. The data stays in your system.
You get seven monitoring screens in your dashboard and the log of every call. The scope and the setup timeline are in writing before we start.
Updated on
You don't need it if your API has little traffic or you already measure each endpoint with its percentiles.
Without measuring every call, the endpoint (each address of the API) that gets optimized is the one someone suspects. With real users, one endpoint takes seconds on one call in a hundred, a screen requests the same thing twice and the average keeps saying everything is fine. With the measurement in front of you, you see which one is slow, which screen calls it and how many people it affects.
In 30 minutes we look at what your API is built with and where you want to store the data. Then you get the scope and the setup timeline in writing.
The capture is installed in your own API and stores the log in your Elasticsearch. Passwords, tokens, emails and locations are masked before they are stored.
You see the seven screens in your dashboard, next to the rest of your console. We review with you, periodically, which endpoint to optimize first.
Priority: High
Endpoint fast on average and slow at p99
An example of the report format, not a client's.
Shall we talk about your case?
A 30-minute meeting to see what you have and what to review first.
All three are useful, each for one thing. AI brings the volume and people bring the judgment: that is why our review uses an automated AI tool and senior developers supervise it.
The review is done by a team of senior developers, with many hours of development and of finding the mistakes AI makes when it codes. We know more than 1,200 review points. The automated AI tool goes through them, and people verify, direct and review.
Team of senior developers
José María Salido, founder of Sondeon, answers for it
It is measuring every call your API receives: how long it takes, what it answers and who makes it, to see where the time goes and what fails before a customer notices.
They are percentiles: p95 is the time that 95% of calls stay under, and p99 is the same for 99%. They show the slow calls the average hides: an endpoint can be fast on average and keep one call in a hundred waiting for seconds.
A Python API built with FastAPI takes it as is; for Node, Django or Spring the capture is adapted. You need Elasticsearch, and a message queue if you already have one.
It does not affect the response: the log entry is built after the response is sent, on another thread, and travels through a queue. The data is stored in your own Elasticsearch, so nothing goes to a third party.
Monitoring tells you which one to optimize first and why, with the measurement in front of you. The fix is separate and only if you ask: your team does it, or we do if you ask us to.
The 30-minute meeting is free. The rest depends on how many APIs are included and on their call volume, and it is settled in writing. It does not replace one: it is a continuous health check that repeats on every call what an API audit would look at, with percentiles and a periodic review of what to optimize first.
Flutter apps built with AI, tested on real iOS and Android phones by command and with touch gestures, together with their backend.
Your website's performance, visibility in search engines and in AI assistants (SEO and GEO), accessibility and security, measured on the real page.
The server side behind your app or your website: what the API exposes, what it protects, what headers it serves and what ends up in its logs.
PostgreSQL, Redis, Elasticsearch and RabbitMQ, in read-only mode: structure, indexes, slow queries, locks and connections.
The server your system will run on, reviewed before you go live: who has access, what runs on it and which limits break as you grow.
The Linux machines your system runs on, measured every few seconds, so you see a full disk or a stopped container before your users do.
Fill in the form and we email you to set a day and a time. It is 30 minutes to see what you have and what to review first.