Notable Changes
- The Databricks CLI is now available on PyPI as
databricks: run it withuvx databricks, or install it withuv tool install databricks,pipx install databricksorpip install databricks. The package contains only the prebuilt CLI binary. Uninstall the legacydatabricks-clipackage first, since both install adatabrickscommand. (#7704, #7777)
CLI
- Add
databricks sandbox fscommands (cp,ls,cat,mkdir,rm) to transfer files between your local filesystem and sandboxes using<sandbox-id>:/path. (#7372) databricks fscommands retry reads (stat, list, open) and deletes on DBFS and Volumes when the server returns a transient error or does not respond within the HTTP timeout. (#7556)- Retry UC Volumes file operations (
fs cp,fs ls,fs rm, bundle uploads to Volumes) when the Files API returns HTTP 408, 429, 502, 503 or 504 (for example "AWS S3 is throttling requests"), or on a transient network error; this regressed in v1.10.0. (#7559) - Stop warning when the host metadata request (
/.well-known/databricks-config) times out; the CLI already falls back to your config, and the message is still shown with--debug. (#7752) - Add
databricks aitools auto-install disableandenableto turn off (or back on) the automatic installation of Databricks AI tools on CLI upgrades. A global-scopedatabricks aitools uninstallalso turns it off. (#7701) - Stop printing the Databricks AI tooling install hint on
auth,configure,current-user me,version,help, and shell-completion commands. (#7452) - When Codex, Cursor, GitHub Copilot CLI, OpenCode, Gemini CLI, Antigravity, Pi, Goose or Kiro run the CLI without Databricks skills installed for that agent, the CLI now prints a one-line hint to install them with
databricks aitools install --agents <agent> --scope global, previously shown only to Claude Code. (#7586) databricks aitools installnow rejects an unknown--agentsname before it shows the scope prompt, instead of after. (#7469)databricks aitools installno longer asks "Proceed?" after the agent picker; the picker selection confirms the plan. The picker title now shows the install scope, its hint says Enter installs,--skills-onlyruns label plugin agents "skills" and pre-check every agent whose config dir exists, and the plan summary after it is reduced to the scope line. (#7478)databricks aitools install --output jsonreports more specificerror_categoryvalues:SKILLS_FETCH_FAILED,RELEASE_LOOKUP_FAILED,RATE_LIMITED,MANIFEST_INVALID,STATE_INVALIDandPLUGIN_CLI_TIMEOUT. A coding agent CLI that does not answer in time is now reported astimed out after <duration>, followed by any stderr it wrote before it was stopped. (#7475)databricks aitools installno longer prompts when a coding agent runs it, even in a pseudo-terminal; it uses the non-interactive defaults instead and prints a note saying so. (#7477)databricks aitools installnow prints the directory it installed skills into (for exampleInstalled 1 skill in ~/.databricks/aitools/skills.), so agents can open the installedSKILL.mdfiles without searching for them. (#7668)databricks aitools installnow ends a failed install with a step to recover, in the text output and in a newrecoveryfield of--output json. (#7480)databricks aitools installnow installs plugins for the other coding agents even when installing skill files fails. (#7479)databricks aitools installnow installs skill files for a detected coding agent whose CLI is not on PATH (for example a Claude Code desktop or IDE install), unless it already has the plugin, says how to switch to the plugin later, and warns when the current project also has the plugin.databricks aitools updateno longer tries to install the plugin for such an agent, and no longer writes skill files next to a plugin that such an agent already has. (#7482)- Fix
databricks aitools installwaiting a long time on an unreachable or slow workspace while it resolves authentication for telemetry; that wait is now limited to a few seconds. (#7468) databricks aitools install,updateandlistnow refuse a skills manifest that names a skill or file outside the skills directory, before writing any skill files. A skill whose name or file paths are not portable is skipped with a warning, andupdatekeeps its installed copy. (#7722)databricks aitools installanddatabricks aitools updaterefresh only the marketplace the Databricks plugin is listed in, so another unreachable marketplace no longer blocks them. When that refresh fails, both commands warn, install or update the plugin from the local copy of the marketplace, and record the plugin version as unknown. (#7473)databricks aitools installanddatabricks aitools updatenow wait up to 20 seconds instead of 5 for a coding agent's CLI to answer the plugin capability check, so a slow first start is less likely to report the agent as unavailable. (#7472)databricks aitools installno longer fails a plugin-only install when the lookup of the latest skills release on GitHub fails or is rate-limited. The plugin is installed and its version is recorded as unknown. (#7474)databricks aitoolscommands now send usage telemetry, and the CLI records when it recommends installing Databricks AI tools. When the configured workspace is slow or unreachable,aitools install,updateanduninstallcan take up to about 6 seconds longer, andlist,versionand runs that stop early up to about 3 seconds. SetDATABRICKS_CLI_DISABLE_TELEMETRY=1to opt out. (#7467, #7469, #7470)databricks aitools installanddatabricks aitools updatenow report a coding agent CLI shim (.cmdor.bat) on Windows that prints its plugin help and then exits with an error asshim-exited-with-error, with the command to run to see the error, instead of reporting the CLI as not on PATH. (#7476)apps initno longer suggests installing skills when every detected coding agent already has the Databricks plugin. (#7467)databricks auth loginnow gives up on the optional workspace lookup after about 5 seconds instead of retrying for about 5 minutes, and explains the fallback to the manual workspace ID prompt. (#7522)- Throttle Claude Code's native plugin install hint separately from the text hint: it prints at most once per 10 minutes per session, so the install dialog still comes up when a subagent, a hook, or a command outside a Claude Code subprocess runs the session's first
databrickscommand. (#7692) - Print Claude Code's native plugin install hint alongside the text hint, so Claude Code offers a one-click install of the Databricks plugin. Claude Code now counts as having the tooling only when it loads it: skills only in
~/.databricks/aitools/skillsdon't count, and a project- or local-scope plugin counts only when the CLI runs inside that project.auth token,auth docker tokenand CLI subprocesses no longer print install hints. (#7585) - Interactive prompts and pagers now work in Git for Windows' bash and other MSYS2 shells when they run under ConPTY (the default since Git for Windows 2.47) or directly in a Windows console. (#7699)
- Add the
databricks feedbackcommand to send feedback about the CLI to Databricks. (#7724) - Fix shell completion in
databricks fscommands for local paths typed with a leading./or repeated separators, which previously offered no suggestions. (#7679) - Fix a panic in
databricks labscommands when a non-string SDK setting such asDATABRICKS_DEBUG_HEADERSorDATABRICKS_RATE_LIMITis set in the environment. (#7673) - Fix a panic in
databricks labs installfor projects whoselabs.ymlhas noinstallsection. (#7673) NO_COLORno longer disables prompts, spinners or the list pager; they run without color. (#7700)- Add the Beta
databricks pipelines testcommand, which runs the tests of a pipeline, and thetestingproperty of pipelines in bundles. (#7429) - Add
databricks sandbox forward, which serves an app running in a sandbox at a local URL. (#7395) - Bound the whole telemetry upload at exit, including config resolution and authentication, by its timeout, so a stalled auth or telemetry endpoint no longer keeps a command from exiting. (#7467)
databricks version --checknow suggestsdatabricks aitools installon stderr when a coding agent is detected without the Databricks plugin or skills. (#7471)databricks aitools install,updateanduninstallnow exit with code 1 when they do nothing they were asked to.installfails when it sets up nothing: no coding agent is detected, every agent is skipped or fails to install, or no skill in the release can be installed (also with--path).updatefails when every update or check it attempts fails, anduninstallwhen every removal it attempts fails. Scripts that ran these commands as a best-effort step and relied on exit code 0 should handle the new exit code. (#7481)- Add
file:<path>selectors topipelines runand pipelinebundle run.--refreshand--full-refreshnow use space-separated lists, such as--refresh "orders payments", instead of comma-separated lists. Single names containing spaces now require backticks, e.g.--refresh 'table:`order details`'. (#7184)
AI Runtime
air get -o jsonnow reportsPENDINGwhile the run waits for compute, matchingair list, and includes the run'sstatus_message. (#7455)- Wrap long failure reasons and log lines in the
air listinfo and logs panes, keep the MLflow link (labeled with its retry) when a run's last retry failed before starting, and pointair logsat earlier retries when the last one has no logs. (#7456) databricks air logs --nodenow reports an out-of-range node index when it can determine the run's node count, instead of looking like a successful run with no logs. (#7516)- Add experiment, duration, accelerator, and dashboard metadata to
air list --output json. For retried runs,started_atnow reports the latest attempt, matchingair get. (#7568) air listshortens MLflow run names and unnamed-run IDs by keeping their end behind a leading…, and its columns no longer overflow on wide (for example CJK) text;air getshows an unnamed run's full MLflow run ID. (#7582)- Shrink the Experiment and User columns of the interactive
air listtable to fit the terminal width, keeping the end of a shortened experiment name. (#7583) - Show the experiment, accelerators and configuration in the
air listinfo pane for runs where it showedN/A. (#7587) - Strip terminal control characters from the MLflow run, experiment and user names shown by
air listandair get, so a run name cannot clear or retitle the terminal of whoever lists it. (#7601) air logs,air get, andair run --watchno longer make an extra credential-check request before fetching the run, andair getreuses the configured workspace ID instead of looking it up. (#7613)air listno longer waits on extra current-user requests before listing runs, which makes the default list about 150ms faster. (#7615)air list --all-statusrequests larger index pages, so it makes far fewer round trips once the server supportspage_size. (#7616)- Fit the interactive
air listtable to narrow terminals: when the Experiment and User columns are at their floors, shorten Started toMM-DD HH:MMand size the MLflow column to its labels, then hide Accelerators, User, MLflow and Progress in that order, naming the hidden columns in the hint line. (#7629) air get -o jsonnow includesmlflow_run_idandmlflow_experiment_id, so a run can be handed to MLflow-based steps without parsingmlflow_url. They identify the MLflow runmlflow_urllinks to, including one from an earlier retry thatmlflow_retrynames. (#7636)air getand theair listinfo pane now show the environment of runs on the Databricks AI environment (databricks_ai_v*) instead ofN/A. (#7637)air runnow rejects acode_source.include_pathspath that selects no files instead of uploading a snapshot without it; glob patterns and pathspec magic are left to git. (#7643)air convert-to-dabsnow suggests the--output-dirthat converts a working-tree workload whosecode_source.root_pathis outside the bundle directory, instead of suggestingsync.paths. (#7644)air runnow lists acode_source.root_pathoutside a git repository relative toroot_path, so running from a subdirectory ofroot_pathno longer uploads an empty or partial snapshot. (#7646)air list --all-statusnow fills--limitpast purged or inaccessible indexed runs. (#7743)air runnow warns about missing PyTorch when a pinned snapshot'sinclude_pathsselects a Python file. (#7744)air cancel JOB_RUN_IDnow rejects run IDs that are not AI Runtime runs instead of cancelling unrelated Jobs runs. In-o jsonoutput, thefailed[].errortext for a run that does not exist or that you cannot view now starts withfailed to get run <id>:instead of the cancel API error. (#7569)air get -o jsonnow reportsmlflow_retrywhen the MLflow link comes from an earlier retry because the run's last retry never reached MLflow. (#7620)databricks air run --helpnow shows how to install the databricks-ai-runtime coding agent skill. (#7652)air run --watch -o jsonnow reports each lifecycleSTATUStransition once and never reports a non-terminal state after the run has ended, including when logs are served from MLflow artifacts. (#7621)air convert-to-dabsnow explains the--output-dirworkaround whencode_source.root_pathresolves to the bundle root, and notes when the generated job carries the defaultmax_retries. (#7579)air runnow refuses to upload acode_sourcesnapshot that selects no files (for example whenroot_pathis git-ignored) and explains the cause, and--dry-runruns the snapshot file selection and reports the file count. (#7570)- Repeat the Jobs run link after live log streaming ends. (#7571)
air get -o jsonincludes run details including flataccelerator_typeandnum_accelerators. Use--include-configto include stored config, which may contain plaintextenv_variablesand parameters; downloads over 1 MiB are omitted, and config-dependentmax_retriesis otherwise null. (#7624)air getand theair listinfo pane no longer showMax Retries 0when the Jobs API leaves the task's retry limit out. For a run submitted withair run, the limit now comes from the run's uploaded config, shown as3 (CLI default)when the config doesn't setmax_retries; it showsN/Awhen the limit can't be known (job runs, multi-container runs, legacy runs with no explicit value, or an unreadable config). (#7623)- Reorganize
databricks airhelp: subcommands are grouped by task,air runhas examples, and YAML fields are listed under headings, with pool fields marked optional andcommand,compute.num_accelerators, andcompute.accelerator_typecorrectly marked required. (#7512) air run -h confignow lists the environment variables AI Runtime sets on every node (such asMLFLOW_RUN_IDandCODE_SOURCE_PATH), andair logs --helpnotes that--download-tocannot be combined with--tailor--minutes. (#7655)- Support full workspace accelerator types in
air listfilters while preserving family matching, and reject invalid filter values. (#7515) databricks air list --all-statusnow warns when the AI Runtime run index is unavailable and it falls back to scanning Jobs runs. That fallback scan is capped at 250 runs, and a stuck index request fails after 15 seconds instead of retrying for minutes.--all-status --filter user=<your email>now uses the run index instead of scanning the workspace's full run history. (#7619)- Return structured JSON errors for invalid
air list --filtervalues when-o jsonis used. (#7612) air list --all-statusno longer fails when a listed run has been purged. (#7523)- Behavior change:
databricks air logsnow prints the logs available now and exits, also for an active run, matchingkubectl logs,docker logsanddatabricks apps logs. Pass--follow(-f) to stream new lines until the run completes;--tail N --followprints the last N lines first.databricks air run --watchstill submits and streams. (#7658) databricks air logs -o jsonanddatabricks air run --watch -o jsonnow emit aWAITINGevent, for example{"type":"WAITING","message":"Waiting for accelerator compute capacity..."}, when the reported waiting status changes. (#7695)databricks air logson a run that ended in the last two minutes now waits for the log index to settle before printing, instead of sometimes dropping lines that were indexed late. (#7663)air logsandair run --watchnow print the waiting status (for exampleWaiting for accelerator compute capacity...) as a line on stderr when stderr is not a terminal, instead of printing nothing until the run produces logs. (#7657)- Route AIR log links to the MLflow Logs page instead of the hidden artifact path. (#7491)
- Invalid
compute.accelerator_typeerrors andair run -h config.computeno longer list the CLI's built-in accelerator types, which could differ from the workspace's. (#7520) databricks air --helpnow notes that GPU pools are not required to run workloads. (#7581)air runnow prints the job run URL with the?o=<workspace-id>suffix, matchingair getandair list;dashboard_urlinair run -o jsonand the--watchevents gains it too. Whenair listcannot resolve the workspace ID, its links now omit?o=instead of ending in?o=0. (#7573)databricks air runnow reports an unknown workload YAML key by its field path, suggests the closest valid field, and points to-h configinstead of naming an internal Go type. (#7574)air run(including--dry-run) now warns before submitting when thecode_sourceimportstorchbut PyTorch is not declared: the environment is the Standard environment (version 5 or later) with notorchdependency, so the run may fail after it waits for a GPU. (#7656)
Bundles
- Add
isolation_modeto thecatalogsresource, so a catalog can be deployed asOPENorISOLATED. (#7463) bundle validateno longer warns when resource permissions and top-level or target permissions name the same principal. (#7291)- Retry bundle file uploads, reads and deletes when the server does not respond within the HTTP timeout ("request timed out after ... of inactivity"). (#7521)
- Show how to recover when
bundle deployfails because an app's service principal does not exist or was deleted. (#7533) - Retry transient server errors when creating Lakebase (
postgres_*) resources in the direct deployment engine. (#7557) - Fix
bundle planreporting thespark_confandcustom_tagsthe backend derives fromis_single_node: trueas drift on clusters. (#7577) - Stop planning a spurious update for Lakeflow Connect pipelines whose
ingestion_definition.data_staging_optionsis filled in by the backend. (#7603) - Retry the root path check in
bundle destroyon transient errors, so a dropped connection no longer fails the destroy. (#7597) - Use the Lakebase branch when its create request times out but the branch was created, instead of failing the deploy, in the direct deployment engine. (#7602)
- Retry transient server errors when deleting resources in the direct deployment engine. (#7708)
- Use the Lakebase project when its create request times out but the project was created, instead of failing the deploy with "already exists", in the direct deployment engine. (#7747)
- Use the feature, model serving endpoint or dashboard that a create request created despite timing out, instead of failing the deploy, in the direct deployment engine. (#7757)
- Use the Lakebase role when its create request fails with a transient error but the role was created, instead of failing the deploy, in the direct deployment engine. (#7767)
- Use the MCP service, model service or model provider service that a create request created despite failing with a timeout, a 499 or a reset stream, instead of failing the deploy, in the direct deployment engine. Create recovery for features, serving endpoints and dashboards now also covers 499 and reset streams. (#7782)
- Retry app updates and deletes in the direct deployment engine when they fail with 400 BAD_REQUEST because the app compute is
UPDATING, after waiting for the compute to settle. (#7606) - Enable drift detection for
source_code_pathandgit_sourceon apps. (#6399) bundle destroyandbundle deployno longer fail with "internal error, missing in state" when the state holds an empty-ID entry left by an interrupted recreate in an older CLI. (#7676)- Fix a feature whose
full_namecontains uppercase letters being recreated on every deploy. (#7398) - Fix
bundle deployment bind(andbundle generate --bind) failing withno such node "resources.internal_immutable_snapshots.immutable"whenworkspace.sync_modeisimmutable, as reported in #7306. (#7329) - Retry creating the deployment lock (without
--force-lock) when the request fails with a transient server error, times out ("request timed out after ... of inactivity") or loses its connection ("unexpected EOF"). (#7594) - Fix a materialization set member whose
feature_namecontains uppercase letters failing to deploy, or recreating the set on every deploy. (#7398) - Delete the old resource when its name follows a renamed schema. (#7718)
- Fix
pipelines dry-runstarting a real run when given a job KEY, and a crash in its shell completion. (#7672) - Rename schemas in place; schemas and volumes follow a renamed catalog or schema. (#7343)
- Fix a panic when validating or deploying jobs with local libraries and serverless job clusters. (#7725)
- Fix
tgzartifacts withgitset producing an empty tarball or failing when the bundle is in a subdirectory of the git repository. (#7674) - Apply
presets.trigger_pause_statusand the development mode pause default totriggers[]entries on jobs. (#7450)