11-05-2020, 02:26 PM
Dear JK,
You state "There is also a limit to how much we can test it ourselves without pu$ng it out onto the platform, so the transition period as we release and harden this technology is going to be a little bumpy - but worth it!"
You fail to understand that the problems far exceed "a little bumpy". JK is very nearing "unusable". If you had a way to monitor when a user could not enter a session or lost connectivity with others or his/her client hung up, etc. you'd be shocked. A great may are seeking other platforms. Many groups are taking a break until things (hopefully) improve.
I understand you may not be able to fully test JK without deploying it "into the wild" but even then there are mitigating strategies:
- Only deploy to a small subset of users. If corresponding server changes are required then build server-affinity into the clients for that subset.
- ACTIVELY MONITOR HOW THE SYSTEM IS BEHAVING.
- Have a plan to rollback changes.
- Keep your users informed.
- Improve your ability to emulate your prod environment in testing.
- Improve your ability to load and pressure test before release.
In order to mature your system and "product" you must develop these capabilities just like any user-facing platform. JK, when working, is a well-liked platform and it seems finally poised to break out. Many of us have seen your monetizing plan but it seems that JK still has quite a way to go before implementing that. Stabilizing your platform, controlling your release processes and monitoring, and ACTIVELY supporting your users should be among your short-term objectives.
You state "There is also a limit to how much we can test it ourselves without pu$ng it out onto the platform, so the transition period as we release and harden this technology is going to be a little bumpy - but worth it!"
You fail to understand that the problems far exceed "a little bumpy". JK is very nearing "unusable". If you had a way to monitor when a user could not enter a session or lost connectivity with others or his/her client hung up, etc. you'd be shocked. A great may are seeking other platforms. Many groups are taking a break until things (hopefully) improve.
I understand you may not be able to fully test JK without deploying it "into the wild" but even then there are mitigating strategies:
- Only deploy to a small subset of users. If corresponding server changes are required then build server-affinity into the clients for that subset.
- ACTIVELY MONITOR HOW THE SYSTEM IS BEHAVING.
- Have a plan to rollback changes.
- Keep your users informed.
- Improve your ability to emulate your prod environment in testing.
- Improve your ability to load and pressure test before release.
In order to mature your system and "product" you must develop these capabilities just like any user-facing platform. JK, when working, is a well-liked platform and it seems finally poised to break out. Many of us have seen your monetizing plan but it seems that JK still has quite a way to go before implementing that. Stabilizing your platform, controlling your release processes and monitoring, and ACTIVELY supporting your users should be among your short-term objectives.