System maintenance

Every three months or so Medallia, Inc. performs routine quarterly maintenance on Medallia Experience Cloud. The type and scope of this maintenance depends on the tasks Medallia needs to complete, which can very quarter to quarter.

When an instance is down for maintenance, services are not available and activities are limited. One week before maintenance is scheduled, and until it is complete, a banner message appears to inform users as follows:

 "Medallia system upgrades will occur on [DATE] between [START TIME and END TIME]. Your program will be intermittently unavailable during this time. If you have any questions, please contact your program administrator for additional information."
Important: Regularly scheduled maintenance happens quarterly, about every three months. Banners go up about one week before the schedule date. For emergency maintenance, Medallia tries to give a half-day's notice, though depending on the severity, the time can be shorter. This topic focuses on scheduled maintenance.

This discussion provides:

  • A general overview of what may happen during maintenance.

  • The impact it can have on companies and programs.

  • Any actionable item that needs to be done by people implementing and servicing each program.

This discussion includes:

Preparing for quarterly maintenance

Medallia notifies program managers and contacts before the maintenance is scheduled. When this happens, the program manager must:

  1. Verify the down-time window for the company instance. Quarterly maintenance usually happens during these times (subject to change):

    • U.K. Servers — Friday 7 p.m. to Saturday 1 a.m. GMT (Friday 11a.m. to 5 p.m. PT).

    • U.S. Servers — Friday 6 p.m. to 12 a.m. PT.

    The windows are scheduled for 6 hours; the actual down-time varies anywhere from 5 minutes to several hours.

  2. Banners usually go up on Friday one week before the scheduled down-time. The same message appears in English to all of the company's users. If the company needs banner in a different language, ontact your Medallia expert.

    Note: There can be only one message — in one language — per program.

  3. Understand the maintenance impact and what will happen during the maintenance period. For detailed information, see Types of maintenance below. Note that one or more types of maintenance can happen during the same period.

  4. Once you understand the impact to your program, communicate the down-time to the company.

  5. Create a checklist to verify everything is running as expected once the system is back up. See After the maintenance completes for areas to check.

    Important: If the program has automated processes that occur during the down-time, it is important to understand how they are impacted (if at all), and to coordinate with the company to align on what will happen.

  6. Notify your colleagues about the down-time, and ensure they understand the impact and the post-maintenance plan.

After the maintenance completes

When the maintenance is complete and Experience Cloud is back up, it is important to verify these areas for proper operations:

  • Feed Pulls triggering as expected
  • Exports triggering as expected
  • Invitations are going out
  • Completed surveys are flowing into the system
  • Users can sign-in to Medallia Web reporting and Medallia Mobile
  • Alerts are triggering

Types of maintenance

During the quarterly maintenance window, Medallia may perform one or more of these types of maintenance: 

Front-end down

AreaStatusDuring maintenance
Medallia Web reporting Down

Users cannot sign-in, and:

  • Some administrators can sign-in, but access should be avoided during this time

  • The maintenance banner appears

Medallia Mobile Down

Users cannot sign-in, and:

  • The maintenance banner appears

  • iOS receives alert notifications

  • Android does not receive notifications

Alerts Depends

Alerts work and escalate per normal. Additionally:

  • If Broadcaster is up, alert emails are sent

  • iOS triggers notifications

  • Android does not trigger notifications

  • Users cannot "close the loop"

APIs Down

APIs are not accessible, and attempts to access them receive an error message. Any API calls made during this time have to be re-issued later when maintenance is complete.

Back-end down

AreaStatusDuring maintenance
Medallia Web reporting Up (data not current)

Users can sign-in, but report data are not current. Additionally:

  • New surveys do not flow into Experience Cloud.

  • Administrators do not have access to Setup.

Medallia Mobile Up (data not current)

Users can sign-in, but report data are not current. Additionally:

  • New surveys do not flow into Experience Cloud.

  • Administrators do not have access to Setup.

Surveys Up (delayed)

Respondents can take surveys. Completed surveys are queued in the Survey engine, and will flow into Experience Cloud when the back-end is back up.

Scheduled Invitation emails (e_invitationdate) to go out during the downtime are queued and will go out when the back-end is back up.

Alerts and reminders Down

Alerts and Reminder emails that would have escalated during  maintenance escalate immediately once the back-end database is back up.

There is no impact on alerts and reminder emails scheduled to escalate after the maintenance period.

Exports and Feed Pulls Down (Coordination may be needed)
Warning: Exports running when the system goes down are immediately cut off, resulting in a partial file. The missing data is not captured in the next export. It is critical to coordinate with companies to align on expectations during the maintenance down-time. Especially verify the file transfer can happen at a different time.

The effect on exports and feed-pulls during maintenance down-time depends on the activity's On Missed Schedule schedule specification. When the Experience Cloud comes back up and the option is:

  • Run Once — The first missed activity immediately triggers, and any other missed activities are skipped. For example, if a schedule runs hourly on the hour, and there was down-time from 2:30 to 5:30, the 3:00 activity is triggered and the next activity happens at 6:00.

  • Run All — All missed activities trigger. For example, if a schedule runs hourly on the hour, and there was down-time from 2:30 to 5:30, each of the missed activities (3:00, 4:00, and 5:00) are triggered, and three separate activities happen immediately.

  • Wait for Next — All missed activities are skipped and the next scheduled activity happens per its schedule.

feed.medallia Queued

Files are received and queued, and when Experience Cloud is back up, they are processed processed in the order received.

Important: Once Experience Cloud is back up and processing the queued files, any new files jump to the front of the queue and are processed before the older files.
Webfeed Down

Files received via Web Feeds queue and are processed in the order received when the back-end is back up.

Tip: If possible, the company should pause sending files during maintenance to limit the impact on shared resources.
Social Media Down

Collectors continue to gather comments and reviews, but the data are queued and flow into when the back-end is back up.

Text Analytics Down

Parsing is suspended and resumes on data pulled prior to the down time.

Omni Exporter Down

Omni Exporter does not run; all scheduled jobs re-trigger once Experience Cloud is back up.

Survey Engine down

The Survey engine manages all surveys that respondents take. In the event a Survey Engine is down, all survey links to that engine will fail, and a browser attempting to link to a survey on the engine may experience a No response or Time-out message. Survey takers can retry and connect to the survey later when the engine is back up. Survey engine usually last no more than 5 minutes, unless otherwise notified.

Broadcaster down

The Broadcaster sends and manages the delivery of emails. When the Broadcaster is down, all emails are queued until the Broadcaster is back up, including (but no limited to):

  • Invitations

  • Reminders

  • Alerts and escalations

  • Rapid responses

  • New user notifications

  • Export successes and failures

  • Feed pull and SFTP successes

  • Processing reports

  • Quota and sampling reports

  • Manual upload confirmations

Feed-server down

The feed servers process file transfers in and out of Experience Cloud. Any process fails to transfer or pull while down, including (but not limited to):

  • Exports

  • Feed Pulls

  • Custom application jobs using the feed servers

You need to re-trigger their process after the server is back up and resend any files or retry any exports.