Telemetry is one of those things that tends to divide a room. On one hand it’s productive and accurate feedback for your product of choice and on the other hand it’s big brother spying on what you’re doing in that product. I wanted to share a recent experience with you based on my recent upgrade of Microsoft System Center Configuration Manager (SCCM) to the recent 1710 release as it may make you change your mind.
The long awaited 1710 release was made available last week a little before midnight in the UK on 20/11/17 (or 11/20 in the US ). The next morning I fired up my lab which runs the current branch version of SCCM, ran the early update ring PowerShell script and proceeded to whizz through the wizard in a speedy not hastily fashion. This was lab after all so on with the day job and I’ll check back on that a bit later. This is of course, for anyone who has experienced the modern SCCM upgrade, the somewhat relaxed approach that Microsoft now afford us – no complaints here. However to my surprise that upgrade failed.
Now being one of those annoying twitter types I tweeted this to my 700 or so followers and said I would hit the logs. The fix here in this post is (at the moment) irrelevant. I remembered that last time I upgraded this environment to 1706 I also did some testing with SQL always on availability groups (AAG) for looking at the passive site server role so I quickly suspended replication on the database and hit the retry. It failed AGAIN. Being that this is after all my lab and I wanted 1710 in there I decided to pull the database out of the AAG altogether. Remember, speed not haste here. 3rd time lucky and we’re in to 1710.
Now, here’s the good bit. Much later that night, just before bed I checked my email. Low and behold I have an email from fellow WMUG leader and MVP Robert Marshal - nothing unusual there – but I’m in the CC line and the mail is to the Director of Engineering for SCCM (David James) along with no fewer than 6 members of the SCCM engineering team. That seemingly insignificant tweet I made earlier in the day had been spotted by one of the team (credit to Mark Silvey here) and shared internally. Little did I think that my failure would be one of only 3 in that first 12 hour period. The team knew this from the telemetry and to my amazement actively followed up through my local MVP (Robert) and contacted me from there. Now I should be clear I’ve met David James before and he knows me through WMUG, he just didn’t have my email address. I don’t imagine he goes hunting down all customers in a similar situation. However, the very fact they knew the failure rates and are actively looking into my failure even after it’s resolved speaks volumes. I was actually 2 of those 3 failures in the first 24 hours. I also picked up on another twitter post from Harjit Dhaliwal and referred the team to that which as it turns out was the 3rd of the 3 failures and I’m happy to say all are now resolved. For the record there were around 300 installs in that first 12 hours and the 1% that failed were soon resolved.
Now I know from David James’ recent visit to WMUG in the UK that he checks the telemetry stats daily and you’ll often see Brad Anderson telling us how many clients are hitting the latest version. It is very clear to me that the information is being used correctly, securely and in the right way. This experience had re-iterated a few things to me though:
I shall leave those thoughts with you and I'd love to hear if anyone has had a change of mind as a result. Finally, to all involved in the SCCM team – keep up the good work.