Dataverse Storage and Capacity: Cutting Cost Without Losing Data Users Need

Which Dataverse capacity bucket is actually full?

Start on the capacity page in the Power Platform admin center. Dataverse capacity is split into three buckets that are pooled across the tenant, and each fills for different reasons, so "storage is full" is not yet a diagnosis. Read the summary for the tenant, then the per-environment view, then open the environment using the most and read its per-table breakdown. That breakdown names the tables growing fastest, which is the list you work from.

Two checks before deleting anything. First, confirm which environment is responsible: sandbox and trial copies of production are often as large as production itself, and retiring an unused copy can recover more than any cleanup. Second, confirm the bucket for each large table in your own breakdown, because Microsoft has changed which system tables count toward which bucket over time.

How do you clean up plug-in trace logs safely?

Plug-in trace logs are the most common surprise because a single setting controls them. In the environment settings, the plug-in trace log option is Off, Exception or All. "All" writes a trace row for every plug-in and custom workflow activity execution, and it is often left on after a debugging session. On a busy environment that setting alone can grow capacity faster than the business data.

Set it to Exception, or Off where nobody reads traces, then remove the history. Recent platform versions include system cleanup of older trace log rows, but do not assume it is running or sufficient in your environment: check the table size in the breakdown and the bulk deletion jobs list.

• Change the setting first, otherwise the table refills while you delete.
• Create a bulk deletion job on the Plug-in Trace Log table with a created-on filter older than the period your developers need for diagnosis.
• Schedule it to recur, so the table stays small without anyone remembering.
• If a developer is mid-investigation, export the traces they need before the job runs.

How do you bulk delete completed system jobs without breaking running processes?

The System Job table (AsyncOperation) records asynchronous plug-ins, classic workflows, bulk deletions, imports, rollup recalculations and other background work. Completed rows stay unless something removes them, so an environment with busy asynchronous logic accumulates them quickly. Microsoft provides bulk deletion job templates for completed system jobs in many environments, but check whether they exist, are active and cover the statuses you expect rather than relying on them.

Filter on status, not only on age. Delete jobs whose status is Completed with a status reason of Succeeded or Canceled. Keep Failed jobs until someone has reviewed why they failed, then delete them on a longer schedule. For classic workflows, the workflow option to automatically delete completed workflow jobs prevents the build-up at source.

• Never include jobs that are Waiting, Waiting for Resources, In Progress, Pausing, Canceling or Suspended. Those are live work, and deleting a waiting workflow removes its pending steps.
• Run the bulk deletion job outside peak hours, since it competes with other background work in the same environment.
• Check the related workflow log rows as well, because they are held in a separate table and can be large.

Which other tables quietly consume database capacity?

After trace logs and system jobs, the per-table breakdown usually shows a longer tail. Each has its own safe route, and each is worth an owner so it does not return.

What should you do about email and note attachments filling file capacity?

File capacity is almost always attachments: files on notes (the annotation table), attachments on email activities, and file or image columns. They matter to users, so this is where cleanup and structural change meet. Deleting old attachments outright is rarely acceptable; moving them out of Dataverse while keeping a link from the record usually is.

Find the sources first: which tables the notes hang off, which mailboxes and tracking rules bring email attachments in, and which processes generate documents such as quotes or signed forms. Then decide per source.

• Stop the inflow: tighten email tracking so signatures, logos and automated notifications are not stored as attachments, and review flows that attach generated files to every record.
• For new documents, use the SharePoint document management integration so files are stored in document libraries linked to the record rather than in notes.
• For existing attachments, move the file content to SharePoint or Azure Blob storage and keep a link or a stub on the record. This needs a migration job, custom code or a tool, tested on a copy first.
• Keep the email activity and its metadata even when the attachment is moved, so the timeline stays complete.
• Verify a sample of moved files opens from the record before deleting the original content from Dataverse.

How do you reduce audit history without losing compliance evidence?

Audit history fills log capacity and is also evidence, so it is the one offender to handle with the people who own that evidence. Two separate decisions apply. Going forward, set the audit retention period for the environment and switch auditing off for tables and columns nobody needs to evidence, such as high-frequency integration fields. For the backlog, audit logs can be deleted from the admin center by age, by table or by type of log, but deletion cannot be undone.

Export before you delete wherever the history may be needed for a regulator, a dispute or an internal investigation, and store the export where it can be searched. What to audit, how to export it and how to prove who changed what are covered in our guide to Dynamics 365 audit trails for compliance. Where users simply need to read recent history on a form, a readable panel such as Solzet Easy Audit makes the retained audit data usable without keeping more of it.

What must never be bulk-deleted from Dataverse?

Bulk deletion is fast and permanent at record level. Environment backups exist, but restoring one restores an environment, not a handful of records, so an overly broad job is expensive to reverse. Keep this list next to every bulk deletion job you create.

• System jobs that are waiting, in progress, suspended or otherwise not completed.
• Records under a retention obligation or a legal hold, including activities and notes relating to a dispute, complaint or investigation.
• Activities, notes and attachments linked to open cases, opportunities or contracts, without first checking the cascade behaviour of the relationship and what the delete removes with it.
• Audit history in a regulated environment before it has been exported and the export checked.
• Staging or integration tables before confirming no integration, report or flow still reads them.
• Anything at all without a verified export or backup and a record of the job criteria, who approved it and when it ran.

What structural changes stop capacity filling up again?

Cleanup buys months; structure keeps the environment inside its entitlement. The aim is that users keep instant access to the history they use, and everything else moves to cheaper storage or leaves on a schedule.

What do retention changes mean for data protection?

Changing how long data is kept is a data protection decision as well as a storage one. Deleting audit history, emails or attachments can remove evidence you are obliged to keep, and keeping them indefinitely, whether in Dataverse, SharePoint, blob storage or an archive, can conflict with retention limits for personal data. Moving data also changes where it is stored and who can access it, which may need recording.

Involve your data protection officer before changing retention settings, bulk deleting personal data or moving attachments to a new store, and record the decision. This page is technical guidance, not legal advice. How retention, erasure and data subject access requests work in Dynamics 365 is covered in our guide to GDPR retention, erasure and DSAR in Dynamics 365.

Is buying more capacity ever the right answer?

Sometimes. If the growth is genuine business data that users need live, and the cleanup and structural changes above are done, additional capacity is a normal cost of a growing system. What should not happen is buying add-on capacity to avoid a cleanup nobody has scheduled; how capacity fits into licence and renewal decisions is covered in our guide to managing Dynamics 365 licensing costs and renewals. A wider review of environments, governance and waste is part of our Dynamics 365 health check and technical audit.

We recommend the right solution - whether that's Microsoft Dynamics 365, Power Platform, or a custom-built CRM. Some businesses need the Microsoft ecosystem. Others need full control without licensing. We deliver both. Where storage and per-user licensing together no longer fit how the organisation works, a custom CRM on React, Node.js, PostgreSQL or .NET gives you direct control of storage and its cost.

How does Solzet help with Dataverse storage and capacity?

We start with the capacity breakdown per environment and table, and a list of the offenders with the space each one holds. Then the safe cleanup: trace log settings, bulk deletion jobs filtered by status and age, with every job criteria recorded and exports taken where evidence is needed. Then the structural work: attachment migration to SharePoint or blob storage, audit retention settings agreed with your DPO, recurring jobs and environment lifecycle, so capacity stays predictable.

Solzet delivers remotely from Yerevan, Armenia, with senior consultants and full-stack developers and 8+ years of Dynamics 365 Customer Engagement and Power Platform work, directly or white-label for Microsoft partners.