Setting up private backups for Sanity using Azure Functions

Exporting the Dataset

The backup function leverages Sanity’s export package to create a zip (tar.gz) file containing the dataset:

async function backup() {
  await exportDataset({
    client: sanityClient,
    dataset: DATASET,
    outputPath: `/tmp/${DATASET}.tar.gz`,
    assetConcurrency: 12,
  });
}

This creates a compressed file of the Sanity dataset in the temporary storage area (/tmp/) on Azure.

Uploading to Azure Storage

Next, the code uploads the backup file to Azure Storage:

const shareName = "shareName";
const directoryName = "backups";
const account = "account";
const accountKey = process.env.AZURE_FILES_ACCOUNT_KEY;

const credential = new StorageSharedKeyCredential(account, accountKey);
const serviceClient = new ShareServiceClient(
  `https://${account}.file.core.windows.net`,
  credential
);

Here, StorageSharedKeyCredential authenticates the Azure Storage client. The client uploads the zip file using a file stream and divides it into manageable chunks.

Automating Backup Cleanup

To avoid storage overflow, we include a function to delete old backups, retaining only the latest 31 files:

async function deleteOldestZips() {
  const directoryClient = serviceClient
    .getShareClient(shareName)
    .getDirectoryClient(directoryName);
  
  let zipFiles = [];
  for await (const item of directoryClient.listFilesAndDirectories()) {
    if (item.kind === "file" && item.name.endsWith(".gz")) {
      const fileClient = directoryClient.getFileClient(item.name);
      const properties = await fileClient.getProperties();
      zipFiles.push({ name: item.name, lastModified: properties.lastModified });
    }
  }

  zipFiles.sort((a, b) => a.lastModified - b.lastModified);
  const filesToDelete = zipFiles.length > 31 ? zipFiles.slice(0, zipFiles.length - 31) : [];

  for (const file of filesToDelete) {
    await directoryClient.getFileClient(file.name).delete();
  }
}

This function ensures only the 31 most recent backup files are retained by sorting the files by modification date and removing the oldest ones.

Testing and Deployment

With the function code ready, we deploy it to Azure Functions, where it now runs automatically each day at midnight. From this moment on, Sanity content backups are reliably stored in Azure, with unnecessary files automatically removed.

Multi-Layered Approach to Data Protection

When it comes to safeguarding critical data, redundancy is key. We strongly recommend enabling backups for your Azure file share as an additional layer of protection against data loss.

The current backup configuration follows a robust retention strategy:

• Daily backups: Created every day and retained for a week to address short-term recovery needs.
• Weekly backups: Generated once a week and stored for a month, providing coverage for medium-term requirements.
• Monthly backups: Conducted once a month and retained for a year to handle long-term data restoration scenarios.

This tiered approach ensures you have multiple recovery points, accommodating a variety of scenarios ranging from accidental deletions to extended outages.

For detailed guidance on how to enable and configure backups for your Azure file share, refer to the official documentation here. Taking these steps will significantly enhance the resilience and recoverability of your data.