Table Grid migrations run in batches on both legs now. Table Grid Migration Tool (TGMT) 4.7 runs several field migrations at the same time on the Table Grid Editor (TGE) to Table Grid Next Generation (TGNG) leg, and the built-in wizard takes multiple grids through one run on the Cloud leg.
I work at Exalate, and our team builds both the Migration Tool and the Cloud migration wizard for Table Grid migration. Batching on both legs currently works well, especially for the recent upgrade of the Table Grid Migration Tool, the mechanism has been proven in practice.
How did migration work before TGMT 4.7
Until TGMT 4.7, the Table Grid Editor to Table Grid Next Generation migration handled one TGE field per run. Select a field, choose a destination, validate, run, read the result, then start again.
On an instance with dozens of grid fields, that meant dozens of separate runs, each started by hand. Plenty of admins kept postponing it.
Leg 1: fields migrate concurrently
TGMT 4.7 lets you select multiple Table Grid Editor fields, configure a destination for each, and run their migrations in one go, without waiting for each field to finish before starting the next.
Each field keeps its own state, progress, issue count, and result. A failure in one doesn't mark the rest as failed, and every result stays attached to its source field.
Each selected field gets its own validation result before you add it to the batch.
Leg 2: grids migrate as a batch
The Cloud migration wizard builds the batch across 3 steps.
Step 3 is grid selection by checkbox. The project count next to each grid shows where that grid is currently used. Clicking the count lists those projects, so you can spot a grid that's still in use somewhere you didn't expect.
Step 4 maps each selected grid to its Cloud destination. By default every selected grid gets the Add new grid option, which creates a grid in Cloud with the same configuration and data as the Data Center one, or you can map to an existing Cloud grid instead.
Step 5 reviews the whole batch before anything runs: number of grids and total warnings.
The result screen reports the outcome as a count, for example Migrated Grids: 4/5, with the failed grids listed and a reason for each. Retry picks up the grids that haven't migrated yet instead of restarting the batch, and the full log downloads from the same screen.
What should you clear before a large batch on leg 2
Each of these affects every grid in the batch at once.
Static rows. If a grid configuration in Data Center contains static rows, those rows get duplicated into every migrated issue in Cloud. Remove them from the configuration before you run, then re-import the configuration with the static data afterwards.
Required-field validation rules. Older grid data may not satisfy the current rules, for example a required column that's empty or a malformed email address. Review doesn't check for these, so the Table Grid Cloud REST API rejects them during the run. Uncheck Required on the affected columns before migrating and re-enable it after.
Unmapped issue types. Issue types in Jira Data Center that aren't associated with any issue type scheme in Cloud get skipped by the Jira Cloud Migration Assistant, and the migration logic rejects them. Create the missing issue types in Cloud, or delete the unused ones in Data Center first.
What should you check before you start a batch
Leg 1: validate every selected TGE field and read its findings, confirm the destination for each, and note the formulas you'll rebuild. Formula results transfer as data. The formulas need rebuilding afterwards.
Leg 2: review the warnings for the selected grids and deal with the 3 issues above. Check Dynamic Options columns separately, because Single and Multi Select List columns using Dynamic Options aren't migrated on this leg. Target columns need to be editable in Cloud before you run, and you can revert that after.
Both legs: rehearse on a sandbox and back up your grid data. On leg 2 specifically, grids that already migrated stay in Cloud and there's no rollback for them.
Table Grid Next Generation is on the Atlassian Marketplace: Table Grid Next Generation.
How many grids are in your largest planned batch, and which of the 3 issues above did you find in them?
