QABot Implementation for ConnectionSphere

[leandrosardi]

Track any error, from any pampa process on the boterrorlog table.

QABot I2P:

• Failed Payments Processing
• IPN Listener is Up
• Unknown IPN records (look for record types who are not included in a list of known values)

QABot Farming List of Checks

• SMSPVA Credits
• Total SignUps and Email Confirmations Last 24 Hours
• Rate of Failed SMS Verification Attempts
• Blocking Rate of New Accounts Last 24 Hours
• Blocking Rate of >0-Invite Accounts Last 24 Hours
• Blocking Rate of >10-Invite Accounts Last 24 Hours
• Blocking Rate of >20-Invite Accounts Last 24 Hours
• Blocking Rate of >30-Invite Accounts Last 24 Hours

QABot Infrastructure

• Central Load Balancer (Ionos/Nginx) is Up
• Central WebServers (Sinatra) are Up
• Divisions Load Balancers (Ionos/Nginx) are Up
• Divisions WebServers (Sinatra) are Up
• Dashboard Exceptions (requires to track errors on pages 500 and 404)
• Processes Exceptions (requires to push to boterrorlog when an error happens on Pampa processes)
• Database Query Timeouts (search for it on boterrorlog)
• Host low resources (from SHM)
• Stuck Processing Threads
• Proxies UpTime (requires to develop the proxies checker job process)

QABot Prospecting Production:

• Accounts assigned to clients must have warming flags OFF
• Number of Available Sales Assistants for New Pipelines
• Rate of Failed Invitation Messages
• Rate of Not Allowed Invitation Messages
• Rate of Failed FollowUp Messages
• Rate of Failed Group Chat Attempts
• Rate of Accepted Connection Requests
• Rate of BotErrorLog / Connection-Requests Delivered , per hour
• Unknown BotErrorLog records (look for record descriptions who are not included in a list of known values)
• Not Edited Operative Accounts
• Blocking of Operative Accounts
• Daily Delivery Quota per Pipeline
• Paused Pipelines with Accounts Assigned
• Pipelines with low rate of new connections
• Pipelines with no answering by neither the client, nor our operators --> should switch it to be handled by operators
• Pipelines with answering but low rate of transfers
• Un-Allocated Expired Resourced
• Not Allocated Resources

QABot Prospecting Sales

• Monthly Subscriptions Volume
• Monthly Subscriptions Growth
• Monthly Clients Retention Rate
• Quarterly Clients Retention Rate
• Yearly Clients Retention Rate
• Customer Success Managers Rate

[leandrosardi]

Track any error, from any pampa process on the boterrorlog table.

Track any error, from any pampa process on the boterrorlog table.

Hierarchy of Classes:

• MyProcess
  • • MyParentProcess
  • • MyChildProcess
  • • • MyLocalProcess --> reporting to boterrorlog
  • • • • MyCrawlProcess
  • • • MyRemoteProcess
  • • • • MyBotProcess --> reporting to boterrorlog

alter table boterrorlog alter column id_lnuser uniqueidentifier null;
go

alter table boterrorlog alter column id_worker uniqueidentifier null;
go

use mirror01;

alter table boterrorlog_history alter column id_lnuser uniqueidentifier null;
go

alter table boterrorlog_history alter column id_worker uniqueidentifier null;
go

[leandrosardi]

After changing the logging, it is neccessary to adapt this operation: https://github.com/leandrosardi/flywheel/issues/610

Also, It is neccessary to replace the comparsion

cast(e.description as varchar(49)) = 'Code: (unknown). Error requesting Mimic to start '

in the source code.