Verbaltech2 - BitShares Witness Report - Monthly Update

October's activities were interesting, especially near the end of the month. I was hoping to officially publish the URLs for 2 load balanced API servers in this report but neither was working last night when I double-checked them. I was also trying to get my BitShares testnet witness back online with the test-2.0.171105 updates so I ran out of time to investigate why the API interfaces were not working. There was also an issue with the witness software containing BSIP18 deployed last week which required rolling back the code on all but one of my nodes to the previous 2.0.171019a release.

I also need to explain why I missed so many blocks last week. I manually check my nodes at least twice a day, first thing in the morning and last thing in the evening. I also have automated notifications setup via email and Telegram. I didn't notice missed blocks until I was on the road, 60 miles away. When I checked in the morning before I left the house the missed block count was 1689 as it was the night before. However, I failed to notice the node I was looking at was no longer in sync. Perhaps that's also why the Telegram bot failed to send me a notification (bot node not in sync?). Needless to say my top priority is to fix notifications and get automatic failover protection in place.

Here is the market summary since last month: (all prices via CMC): BTC=$7287.04, BTS=$0.059, STEEM=$0.88, PPY=$3.77.

---

This month required updates to all nodes several times. Thought load balanced APIs were good to go but not so, I have more to investigate on that. That will be lowest priority, coming in after resolving the issue with notifications and implementing a failover mechanism in python as I mentioned last month.

The testnet server is now running the test-2.0.171105 release and humming along to assist with further testing of the failed production code that contained BSIP 18. Not sure how much time witnesses put in on testing BSIP 18, and witness skills also vary, perhaps widely. As I wrote last month the testing model Peter relied on imposes significant requirements on witnesses to understand C++ programming as well as the economics around black swan events. Some witnesses may not be programmers and others not much into economics. Bottom line is the approach taken may not have resulted in the best test coverage. Hard to say for sure, we may know more once the bug(s) or algorithm deficiencies are discovered.

That's my report this month. I decided there's no good reason not to show the graphs for the actual block producer, so I will also include those graphs from now on.

France

Romania

Germany

Japan

Block Producer (witness) node:

And the testnet server for BTS:

Your vote for witness
Verbaltech2
is greatly appreciated! Thanks for your time and attention