Why building Corners?
SaaS Monitoring
The monitoring landscape Metrics, Logs and Traces. You track resources consumption with Metrics. You use Traces to follow request across your server. Logs to understand how the code interacts.
When you start you are probably in a serverless landscape only logging is available and maybe some metrics are available but it’s slim.
You also have analytics and you can do events in analytics, setup events on buttons and in the back-ends.
In more advanced scenario you have a prometheus and grafana instances you can make custom metrics and dashboards but still it’s not exactly what I am looking for.
Now
- Monitoring without the right timing is 10x less effective. A log loses its value as time goes
- Your logging system, most of the time, does not have all the data
- Metrics are nice but a late game addition
- Each of the team member will try to get a picture with the data that they have.
- Most of the time, the customer message you or open a ticket then you react to an anomaly
The Problem
It’s hard to get an overview of your entire system using only logs, metrics and analytics. It’s also hard to drill-down into one user or one organization.
Find the events that brought them to an upgrade? Is there a patterns that lead to upgrade; “all organizations that upgraded made use of feature A”
Did they encounter too many errors and churn? Is there a pattern there; a buggy form a logout problem, few features that presents a lot of bugs?
Reconstructing Pattern Is Hard So Imagine Acting On Them
At the past, your developer would need to: extract the logs for an organization/customer, discard and clean the uninteresting one, then present the logs to the rest of the team.
The team would add their inputs and could assign someone to do an action. Still you would only have a picture for one organization and the work would have to be done again for another entity.
Timing is also crucial. Building a filtered collection of logs and assembling the team takes time.
Then deciding on what to do with a situation. By the time the tasks are done could the customer already have churned?
The go-to-market
what’s working, what’s not?
Which action brought high quality leads? How far those leads were able to go in your funnel?
Did we have to pick up the ball mid way because feature Z broke? or we are missing an instruction
Corners
First, I kind of always wanted an event based monitoring platform. I can’t explain why it always felt like it would be nice.
Make tools for yourself first
I’ve been in 100+ people team, but also on very small team and in both case I could see this platform fitting.
Monitoring Is A Team Effort
An event gets in, someone react or comment the team can give inputs or tag someone.
External systems, payment systems, identity, ticket & support, CRM and your own stack in one panorama; a wide-angle view.
Option to filter notifications, forward to sagents
You can pick one entity and get the full story by looking at all the events. You can detect pattern in the usage of your SaaS by comparing churned or successful customer