ExeWatch: 2,000 companies, 74,000 devices and the MVP program

More than 2,000 companies and over 74,000 monitored devices in a few months, the EW-MVP program starting from Brazil with Sileide Campos, and the Developer Guide many of you asked for.

TL;DR: ExeWatch has passed 2,000 companies and 74,000 monitored devices. The EW-MVP program starts, one contact point per macro-region of the world: the first is Sileide Campos for Brazil, and applications for the other areas are open at exewatch@bittime.it. And the Developer Guide is online, the thing you asked for more than anything else. Try it at exewatch.com.

The numbers

More than 2,000 end customer companies run ExeWatch in production, on over 74,000 devices. In a few months.

We did not expect them this soon. The reason they keep coming is always the same though: if you write Delphi software you need to know how your applications behave once they are out the door, and until yesterday you made do with the logs you asked the customer to email you, with phone calls and, on a good day, with a crash report. A crash report tells you where the application died, not what it was doing the minute before, how many times it had already happened, and to how many customers.

Moving from reconstructing the facts afterwards to proactive monitoring is the whole difference. The web world has taken application telemetry for granted for fifteen years. The desktop and client/server world, which is where the business software that companies invoice with actually runs, has not.

The EW-MVP program

An EW-MVP is a person who actually uses ExeWatch and knows it inside out, and who is the reference point for their area: they answer questions in the local language, show the product to whoever is evaluating it, and bring back to us what developers in their area need.

It exists for a practical reason. ExeWatch is used in countries whose language we do not speak and whose industrial fabric we do not know. Above all, in those countries at three in the afternoon local time we are not even awake.

The plan is to cover every macro-region of the world, one at a time, with an EW-MVP of reference.

The first EW-MVP: Sileide Campos, Brazil

Our first EW-MVP is Sileide Campos, for Brazil. She started telling Brazilian developers about ExeWatch on September 1st, from the stage of the Embarcadero Conference.

Sileide is also an Embarcadero MVP, so explaining tools to developers is what she does for a living. With her the conversation does not start from the feature list, it starts from what your company needs: which metrics to watch, what to instrument first, which ExeWatch features change your day and which ones might matter less to you.

We started there because the Brazilian one is among the liveliest Delphi communities on the planet.

If you develop in Brazil and you want to stop guessing how your applications are doing and start knowing, write to Sileide: sileidecampos@gmail.com. She will explain in Portuguese what ExeWatch can do for you and for your company, and you will not have to wait for us to wake up in Rome.

Everybody else, for now, can keep writing directly to exewatch@bittime.it.

Apply as an EW-MVP

If you like ExeWatch, you know it well and you think you can be the reference point for your country or for a macro-area, write to us at exewatch@bittime.it.

What we look for:

• that you already have an audience in your area. A user group, a channel, a blog, the local conferences, the group where people come to ask you things;
• that you actually want to talk about it. Talks, demos, articles, answers in the local groups. We are not asking for fixed days, we are asking that you want to do it;
• that you speak the local language. A question answered in the language of the person asking it, inside their own time zone, gets closed in an afternoon.

What you get:

• a direct line to us. What you report, and what your users report to you, goes into the roadmap. You see the betas first;
• visibility. You are the official reference for your area, with your name and contact details on the ExeWatch site and in posts like this one. That is exactly why you are reading Sileide’s name;
• money. The role is a technical one, but it is not volunteering: on the customers you bring in there are terms we agree on together, case by case.

Write to us and tell us where you are, what you build, and how you use ExeWatch today.

The Developer Guide

Of all the things you have asked us for in these months, this one came up more often than any other: a guide that tells you what to instrument, not what the functions are called. I wrote it, and here it is: www.exewatch.com/devguide.

Sixteen sections. The API reference was already there, this one answers the question that comes before it: what do I instrument, and at what level. Inside you will find:

• the quickstart, which is two lines of code and one configuration option;
• the integration of every SDK: Delphi, Free Pascal, .NET, Python, JavaScript and the flat C ABI DLL you call from C++Builder, Microsoft Visual C++ and Visual Basic;
• the five questions, the map between the problem in front of you and the tool that solves it;
• eight recipes on real cases, like “a customer says it crashed and you have no idea why” or “the app flies at 9am and crawls at 5pm”;
• counter, gauge or timing: which of the three to pick, and on what criterion;
• the alerts that are worth having and the ones you mute after three days;
• comparing versions with the CSV export, to catch the regression you carried over from the last release;
• the anti-patterns, which is the part I would read first;
• what happens when the internet connection drops;
• how to instrument without burning through your quota.

If what you want is just the snippet to paste, they are all in Steal this code.

If you already have ExeWatch in production, read at least the anti-patterns and the section on tags. Twenty minutes that save you a dashboard full of stuff you will never look at.

Thanks, and keep it coming

Some of the messages that reached us in these months:

Quotes in the language each person wrote them in, translation under each one. Initials instead of names.

Thirty bugs found and fixed in the first weekend, across 160 devices. That is the kind of report that tells us where to go next.

The direction ExeWatch is moving in was not decided by us in a meeting. Nested Timing Traces and support for legacy C++ and MSVC applications came from an email by somebody who was using ExeWatch and told us what they were missing.

The next one is multi-tenant inside a single process: more than one customer asked for it, wanting separate traces out of a single executable, split by customer or tenant. We will write about it soon.

So keep it coming: write to us when something goes wrong, and write to us above all when something is missing.

If you have not tried ExeWatch yet, the Hobby plan is free and asks for no credit card: start from exewatch.com and then open the Developer Guide.