Connected Solutions Ideas Portal
The modules that are integrated with the Alert Management module (including THM, EQM, etc) as well as DOC currently have limited controls to prevent large numbers of text / email notifications from being sent, risking PwC platform performance, stability (reputation scores in Twilio/Sendgrid), and cost, as well as client user experience:
On AM:
Client roles have access to set up alerts without PwC / Partner Super Admin supervision/control
Alert rules allow setting up alerts that might not be practical (eg. 0 degree alert threshold for temp sensor that is placed in a location that will always be above 0 degrees)
Alerts are retriggered every time a message comes through if the device / alert criteria are met, regardless of whether there's a prior active alert for that same device/alert rule
Devices may send messages very frequently (sometimes several an hour or more), that could trigger alerts
There aren't limits to the number of devices / users that can be included in Alert Groups / On-Call groups, or "uniqueness" enforced for emails / phone numbers
TBD Can alerts/alert groups/on call groups be duplicated? (i.e. same device is part of multiple alert groups that notify multiple times for the same device message)
TBD: If an email / phone number is invalid, no longer active/able to receive messages, or blocks our number, we still send the notification and assume the cost (client doesn't know this is happening)
DOC:
Same as AM, though there are fewer "criteria" clients are allowed to adjust for alert rules (open, close, tampered)
RRB:
RRB doesn't alert On Call users if there is a prior active notification from the same device / location in sequence (different locations for same device should "reset" this). However RRB doesn't have any limit to On-Call users or total text / email usage like the other apps, nor does it enforce unique emails/phone numbers in the on-call list (creating duplicate message risks)
Other considerations:
Sendgrid emails and Twilio texts are about $.01 per message. (Need to verify this and any message limits)
The RRB functionality where an active alert from the same device/location in sequence doesn't trigger a new alert may not be the best solution. Alerts that go unacknowledged for days/weeks/longer could "block" new alerts from triggering notifications. A "cooldown" or other type of alerting criteria may be a better way to enhance this
The alerting functionality may want to consider situations where a situation is getting worse (eg. a water leak that is escalating, a temperature that is rising above more critical / severe thresholds, etc). These may warrant a new alert notification that would override other "cooldown" criteria (or maybe we allow different alerts for the same device groups for this reason)
Is there a "Max" notification volume we want to enforce? If so, is this a function of the number of devices? Other criteria?
We don't pass on alert notification costs, show clients the potential cost of an alert rule (azure alerts are a great example of showing this), show clients if there's a problem with an email/phone number, etc. Do we want to attempt to do this?
Suggest evaluating short term / medium term fixes:
Short term - can we set up alerts in Twilio/Sendgrid of abnormal usage so PwC can respond to a situation?
Medium term - are ther practical changes that can be made to the above to prevent the issues above, while still providing a great customer experience? Can these be made across the modules (including DOC / RRB)?
