[{"data":1,"prerenderedAt":807},["ShallowReactive",2],{"docs:en:\u002Fdocs\u002Fworkflows\u002Fruns":3},{"page":4,"navigation":680,"surround":802,"prefix":683},{"id":5,"title":6,"body":7,"description":672,"extension":673,"meta":674,"navigation":675,"path":676,"seo":677,"stem":678,"__hash__":679},"agentDocs\u002Fdocs\u002F4.workflows\u002F6.runs.md","Runs",{"type":8,"value":9,"toc":658},"minimark",[10,14,19,22,47,50,53,57,72,75,78,82,100,106,110,117,124,131,137,155,158,171,182,186,293,296,310,313,317,329,332,338,342,345,372,379,383,386,424,427,430,435,442,446,452,541,548,563,567,570,615,638,655],[11,12,13],"p",{},"A run is a single execution of a workflow. It runs on the server. A closed tab, a page reload or a server restart does not end it.",[15,16,18],"h2",{"id":17},"start-a-run","Start a run",[11,20,21],{},"You start a run in one of these ways:",[23,24,25,34,44],"ul",{},[26,27,28,29,33],"li",{},"in the tab ",[30,31,32],"strong",{},"Run"," of a published workflow that can be started by hand",[26,35,36,37,40,41],{},"with ",[30,38,39],{},"Run once now"," on a workflow with the trigger ",[30,42,43],{},"Schedule",[26,45,46],{},"as a test run in the builder",[11,48,49],{},"The trigger also starts runs by itself.",[11,51,52],{},"If more than eight runs are running at the same time, the others wait visibly in the queue. Runs started by hand get their turn before runs started automatically, such as runs from emails or a schedule. This also applies to a run whose approval you have just answered.",[15,54,56],{"id":55},"follow-a-run","Follow a run",[58,59,60,69],"ol",{},[26,61,62,63,65,66,68],{},"Open ",[30,64,6],{}," in the sidebar, or the section ",[30,67,6],{}," on the overview of a workflow.",[26,70,71],{},"Click the row of the run.",[11,73,74],{},"The timeline of the run opens as a panel over the workflow overview. You can share its address. As long as the run is running, new steps appear automatically.",[11,76,77],{},"The timeline shows the start and end of each step with duration and tokens. It also shows tool calls with arguments and response, context fields that were entered, branches, questions, approvals, errors, interruptions and the end. Steps are collapsed. A click on a step shows its tool calls.",[15,79,81],{"id":80},"stop-a-run","Stop a run",[11,83,84,85,88,89,91,92,95,96,99],{},"Click ",[30,86,87],{},"Stop",". The button is in the run panel of the tab ",[30,90,32],{}," and, during a test run, on the trigger card in the builder. In the builder, ",[30,93,94],{},"Restart"," also ends the current run. The run ends with the status ",[30,97,98],{},"Cancelled",".",[11,101,102,103,105],{},"There is no stop button in ",[30,104,6],{},". You cannot stop another person's run, because you do not see it.",[15,107,109],{"id":108},"resume-an-interrupted-run","Resume an interrupted run",[11,111,112,113,116],{},"A run with the status ",[30,114,115],{},"Interrupted"," is not finished. For a temporary disruption, lowcloud resumes it by itself. Otherwise it waits for a person.",[11,118,119,120,123],{},"If the AI service is overloaded, disrupted or cannot be reached, lowcloud tries again after 1, 5 and 15 minutes. The panel then shows ",[30,121,122],{},"Interrupted, will be resumed automatically"," with a countdown and has no buttons. lowcloud retries a timeout once after 1 minute. If the step has tools that write without approval, the run instead waits for a person immediately, because something may already have happened when it broke off. A run is resumed automatically at most five times.",[11,125,126,127,130],{},"For other disruptions, the panel shows ",[30,128,129],{},"Run interrupted"," and names the reason. The reasons are invalid access, used-up credit, a used-up monthly limit, a missing account, exhausted automatic retries, or a server restart during a run started by hand.",[132,133,134],"warning",{},[11,135,136],{},"After an interruption at a step boundary, actions may have been executed twice. Before resuming, check in the timeline what has already run.",[58,138,139,149],{},[26,140,141,142,145,146,148],{},"Open the run. The button ",[30,143,144],{},"Resume"," in ",[30,147,6],{}," only opens it, it does not resume it.",[26,150,84,151,154],{},[30,152,153],{},"Resume run"," in the panel.",[11,156,157],{},"lowcloud first checks the credit of the workspace and the monthly limit of the person. The run continues from the last saved point.",[11,159,160,161,164,165,168,169,99],{},"If you do not want to resume the run, click ",[30,162,163],{},"Discard run"," and confirm with ",[30,166,167],{},"Yes, discard",". The run then ends with the status ",[30,170,98],{},[11,172,173,174,176,177,99],{},"On the entry ",[30,175,6],{}," in the sidebar, a yellow badge shows how many runs are waiting for a person. When you get an email about this is described under ",[178,179,181],"a",{"href":180},"\u002Fdocs\u002Fapprovals","Approvals and questions",[15,183,185],{"id":184},"run-statuses","Run statuses",[187,188,189,202],"table",{},[190,191,192],"thead",{},[193,194,195,199],"tr",{},[196,197,198],"th",{},"Status",[196,200,201],{},"Meaning",[203,204,205,216,226,239,248,258,268,283],"tbody",{},[193,206,207,213],{},[208,209,210],"td",{},[30,211,212],{},"Waiting to start",[208,214,215],{},"The run is queued. Nothing has been executed yet.",[193,217,218,223],{},[208,219,220],{},[30,221,222],{},"Running",[208,224,225],{},"The run is being executed.",[193,227,228,233],{},[208,229,230],{},[30,231,232],{},"Waiting for a response",[208,234,235,236,99],{},"The run is stopped at an ",[178,237,238],{"href":180},"approval or a question",[193,240,241,245],{},[208,242,243],{},[30,244,115],{},[208,246,247],{},"A disruption stopped the run. It is resumed automatically or waits for a click.",[193,249,250,255],{},[208,251,252],{},[30,253,254],{},"Successful",[208,256,257],{},"The run completed.",[193,259,260,265],{},[208,261,262],{},[30,263,264],{},"Failed",[208,266,267],{},"The run ended with an error. It cannot be resumed.",[193,269,270,274],{},[208,271,272],{},[30,273,98],{},[208,275,276,277,279,280,282],{},"A person ended the run: with ",[30,278,87],{},", ",[30,281,163],{}," or a rejection without a reason.",[193,284,285,290],{},[208,286,287],{},[30,288,289],{},"Skipped",[208,291,292],{},"The first step classified the event from the trigger as not its responsibility. This is not an error.",[11,294,295],{},"A run fails when the agent declares the task impossible, a step reaches its pass limit, a promised context field stays empty, or an intermediate state cannot be saved. The timeline shows the reason in a red line. A banner on the workflow overview says that the last run failed.",[11,297,298,299,301,302,304,305,307,308,99],{},"For email workflows, ",[30,300,289],{}," is the most frequent outcome. In the section ",[30,303,6],{}," of the workflow overview, such runs are hidden by default, and they do not count towards the success rate. In ",[30,306,6],{}," in the sidebar, you find them via the filter ",[30,309,289],{},[11,311,312],{},"A run can also end without having run, for example when the trigger was changed and an older event is no longer processed. The reason is then in the answer line of the run.",[15,314,316],{"id":315},"runs-list","Runs list",[11,318,319,321,322,325,326,99],{},[30,320,6],{}," in the sidebar shows the latest runs across all workflows, newest first. The columns are status, workflow, input, time and duration. Below the workflow name are the run number and the version, for example #12 · v3, or ",[30,323,324],{},"Draft",". A run through your own setup carries the badge ",[30,327,328],{},"You",[11,330,331],{},"Above the list you filter by status and search by workflow or input. Only statuses that occur in the list appear as filters, each with a count.",[11,333,334,335,337],{},"The list shows at most 50 runs. Filters and search work only on these 50. There is no paging and no export. The complete retained history of a workflow is in the section ",[30,336,6],{}," of its overview. There you filter by version and status.",[15,339,341],{"id":340},"run-result","Run result",[11,343,344],{},"Below the timeline of a finished run is what came out:",[23,346,347,353,363],{},[26,348,349,352],{},[30,350,351],{},"Context",": one line per field with the value and the step that entered it. For an empty field, the reason is shown. If fields stayed empty, a notice above the card leads to the affected step.",[26,354,355,358,359,362],{},[30,356,357],{},"Result",": the answer of the run. ",[30,360,361],{},"Show raw text"," shows it as raw text.",[26,364,365,368,369],{},[30,366,367],{},"Questions about the run",": a chat that answers questions about this run from the stored history. It changes nothing and gives no recommendation for a decision. It offers starter questions, for example ",[30,370,371],{},"Why did the agent act this way?",[11,373,374,375,378],{},"In the builder, ",[30,376,377],{},"Replay"," marks the course of the run step by step on the canvas. There you can also repeat a single step of a run with the current state of the canvas. This works for test runs and for runs of your setup. The trigger cannot be repeated this way.",[15,380,382],{"id":381},"retention","Retention",[11,384,385],{},"lowcloud deletes runs according to these rules:",[187,387,388,398],{},[190,389,390],{},[193,391,392,395],{},[196,393,394],{},"Rule",[196,396,397],{},"Limit",[203,399,400,408,416],{},[193,401,402,405],{},[208,403,404],{},"Age of a run",[208,406,407],{},"30 days",[193,409,410,413],{},[208,411,412],{},"Age of a skipped run",[208,414,415],{},"7 days",[193,417,418,421],{},[208,419,420],{},"Runs per workflow",[208,422,423],{},"the newest 30 remain, skipped ones do not count",[11,425,426],{},"Runs in which someone granted or rejected an approval are exempt from the age limit. The limit of 30 runs per workflow still applies to them. lowcloud never deletes runs that have not ended.",[11,428,429],{},"Usage and costs stay in the billing, even when a run is deleted. You cannot configure the retention. You cannot delete individual runs either. Only a lowcloud platform administrator can do that.",[132,431,432],{},[11,433,434],{},"Open approvals and questions expire after 30 days, interrupted runs after 7 days. The interface shows no countdown for this.",[11,436,437,438,441],{},"When an approval or question expires, lowcloud ends the run. The action of an expired approval is not executed. This also applies when the page ",[30,439,440],{},"Approvals"," shows \"As long as it takes\".",[15,443,445],{"id":444},"dashboard","Dashboard",[11,447,448,449,451],{},"The ",[30,450,445],{}," at the top of the sidebar shows spend, use and open items of the workspace.",[187,453,454,464],{},[190,455,456],{},[193,457,458,461],{},[196,459,460],{},"Tab",[196,462,463],{},"Content",[203,465,466,491,512,522,532],{},[193,467,468,473],{},[208,469,470],{},[30,471,472],{},"Overview",[208,474,475,476,279,479,482,483,486,487,490],{},"The key figures ",[30,477,478],{},"Members",[30,480,481],{},"Usage"," and ",[30,484,485],{},"Credit left",". Below them ",[30,488,489],{},"Open items"," with counts per workflow, for example open approvals and questions, stopped and failed runs.",[193,492,493,497],{},[208,494,495],{},[30,496,481],{},[208,498,499,500,279,503,279,506,482,509,99],{},"Billed amounts in the chosen period, split into ",[30,501,502],{},"Workflows, including test runs",[30,504,505],{},"Chat",[30,507,508],{},"Build",[30,510,511],{},"Knowledge",[193,513,514,519],{},[208,515,516],{},[30,517,518],{},"Agents",[208,520,521],{},"Per agent and workflow: use with the number of failures, people, spend and last use.",[193,523,524,529],{},[208,525,526],{},[30,527,528],{},"History",[208,530,531],{},"One entry per chat message sent and per run: time, person, agent or workflow, and for a run also the status. The tab shows no content.",[193,533,534,538],{},[208,535,536],{},[30,537,478],{},[208,539,540],{},"Per member: last active, runs, chats and spend.",[11,542,543,544,547],{},"In a client workspace, the tab ",[30,545,546],{},"Prices"," is added.",[11,549,550,551,279,553,279,555,482,557,559,560,562],{},"You choose the period from fixed spans or as a range of your own in the calendar. It applies to the tabs ",[30,552,481],{},[30,554,518],{},[30,556,528],{},[30,558,478],{},". The ",[30,561,472],{}," is not limited to a period. There is no export.",[15,564,566],{"id":565},"permissions","Permissions",[11,568,569],{},"Who sees and can answer a run depends on whose accounts it runs through. New runs are always runs of a personal setup or test runs. Runs through the accounts of the workspace exist only as old runs.",[187,571,572,581],{},[190,573,574],{},[193,575,576,578],{},[196,577,32],{},[196,579,580],{},"Visible and answerable for",[203,582,583,591,599,607],{},[193,584,585,588],{},[208,586,587],{},"Run of a personal setup",[208,589,590],{},"only the person whose accounts it runs through",[193,592,593,596],{},[208,594,595],{},"Test run in the builder",[208,597,598],{},"only the person who started it",[193,600,601,604],{},[208,602,603],{},"Old run through the workspace accounts, workflow visible to the team",[208,605,606],{},"every member of the workspace who sees the workflow",[193,608,609,612],{},[208,610,611],{},"Old run through the workspace accounts, workflow private",[208,613,614],{},"only the author",[11,616,617,618,620,621,623,624,627,628,631,632,635,636,99],{},"Other people's runs do not appear in ",[30,619,6],{},", in ",[30,622,440],{}," or as a run detail. You cannot delegate or forward an approval to another person. You see other people's runs only as numbers without content. Anyone who may build (the role ",[30,625,626],{},"admin",", or the former role ",[30,629,630],{},"developer",") sees them per person in the tab ",[30,633,634],{},"Setups"," of the workflow. Admins see them in the ",[30,637,445],{},[11,639,640,641,643,644,647,648,651,652,99],{},"Only admins open the ",[30,642,445],{},". lowcloud redirects everyone else to the workflow list. People with the role ",[30,645,646],{},"member"," do not see a credit amount. Instead, they see the card ",[30,649,650],{},"Your monthly limit"," with their own spend in the ",[30,653,654],{},"Settings",[11,656,657],{},"Only someone who may build can repeat a step of a run.",{"title":659,"searchDepth":660,"depth":660,"links":661},"",2,[662,663,664,665,666,667,668,669,670,671],{"id":17,"depth":660,"text":18},{"id":55,"depth":660,"text":56},{"id":80,"depth":660,"text":81},{"id":108,"depth":660,"text":109},{"id":184,"depth":660,"text":185},{"id":315,"depth":660,"text":316},{"id":340,"depth":660,"text":341},{"id":381,"depth":660,"text":382},{"id":444,"depth":660,"text":445},{"id":565,"depth":660,"text":566},"Follow runs, stop them and resume interrupted runs.","md",{},true,"\u002Fdocs\u002Fworkflows\u002Fruns",{"title":6,"description":672},"docs\u002F4.workflows\u002F6.runs","Xxb0yK7gQ5t75HW5t2ssL81SyOfPAzNhvwdoBO7hqZQ",[681],{"title":682,"path":683,"stem":684,"children":685,"page":-1},"Docs","\u002Fdocs","docs",[686,689,707,711,735,762,765,779,784],{"title":687,"path":683,"stem":688},"Documentation","docs\u002Findex",{"title":690,"icon":691,"path":692,"stem":693,"children":694,"page":706},"Getting started","i-lucide-compass","\u002Fdocs\u002Fgetting-started","docs\u002F1.getting-started",[695,698,702],{"title":472,"path":696,"stem":697},"\u002Fdocs\u002Fgetting-started\u002Foverview","docs\u002F1.getting-started\u002F1.overview",{"title":699,"path":700,"stem":701},"Quickstart","\u002Fdocs\u002Fgetting-started\u002Fquickstart","docs\u002F1.getting-started\u002F2.quickstart",{"title":703,"path":704,"stem":705},"Glossary","\u002Fdocs\u002Fgetting-started\u002Fglossary","docs\u002F1.getting-started\u002F3.glossary",false,{"title":505,"path":708,"stem":709,"icon":710},"\u002Fdocs\u002Fchat","docs\u002F2.chat","i-lucide-message-circle",{"title":518,"icon":712,"path":713,"stem":714,"children":715,"page":706},"i-lucide-bot","\u002Fdocs\u002Fagents","docs\u002F3.agents",[716,720,723,727,731],{"title":717,"path":718,"stem":719},"Building agents","\u002Fdocs\u002Fagents\u002Fbuilding","docs\u002F3.agents\u002F1.building",{"title":511,"path":721,"stem":722},"\u002Fdocs\u002Fagents\u002Fknowledge","docs\u002F3.agents\u002F2.knowledge",{"title":724,"path":725,"stem":726},"Skills","\u002Fdocs\u002Fagents\u002Fskills","docs\u002F3.agents\u002F3.skills",{"title":728,"path":729,"stem":730},"Sharing and access","\u002Fdocs\u002Fagents\u002Fsharing","docs\u002F3.agents\u002F4.sharing",{"title":732,"path":733,"stem":734},"Microsoft Teams","\u002Fdocs\u002Fagents\u002Fteams","docs\u002F3.agents\u002F5.teams",{"title":736,"icon":737,"path":738,"stem":739,"children":740,"page":706},"Workflows","i-lucide-workflow","\u002Fdocs\u002Fworkflows","docs\u002F4.workflows",[741,745,749,753,757,761],{"title":742,"path":743,"stem":744},"Your first workflow","\u002Fdocs\u002Fworkflows\u002Ffirst-workflow","docs\u002F4.workflows\u002F1.first-workflow",{"title":746,"path":747,"stem":748},"The builder","\u002Fdocs\u002Fworkflows\u002Fbuilder","docs\u002F4.workflows\u002F2.builder",{"title":750,"path":751,"stem":752},"Steps","\u002Fdocs\u002Fworkflows\u002Fsteps","docs\u002F4.workflows\u002F3.steps",{"title":754,"path":755,"stem":756},"Triggers","\u002Fdocs\u002Fworkflows\u002Ftriggers","docs\u002F4.workflows\u002F4.triggers",{"title":758,"path":759,"stem":760},"Publishing and setup","\u002Fdocs\u002Fworkflows\u002Fpublishing","docs\u002F4.workflows\u002F5.publishing",{"title":6,"path":676,"stem":678},{"title":440,"path":180,"stem":763,"icon":764},"docs\u002F5.approvals","i-lucide-check-check",{"title":766,"icon":767,"path":768,"stem":769,"children":770,"page":706},"Integrations","i-lucide-cable","\u002Fdocs\u002Fconnections","docs\u002F6.connections",[771,775],{"title":772,"path":773,"stem":774},"Connecting services","\u002Fdocs\u002Fconnections\u002Fservices","docs\u002F6.connections\u002F1.services",{"title":776,"path":777,"stem":778},"Custom integrations and MCP","\u002Fdocs\u002Fconnections\u002Fcustom","docs\u002F6.connections\u002F2.custom",{"title":780,"path":781,"stem":782,"icon":783},"Sovereignty","\u002Fdocs\u002Fsovereignty","docs\u002F7.sovereignty","i-lucide-shield-check",{"title":785,"icon":786,"path":787,"stem":788,"children":789,"page":706},"Workspace","i-lucide-users","\u002Fdocs\u002Fworkspace","docs\u002F8.workspace",[790,794,798],{"title":791,"path":792,"stem":793},"Workspace and roles","\u002Fdocs\u002Fworkspace\u002Froles","docs\u002F8.workspace\u002F1.roles",{"title":795,"path":796,"stem":797},"Credit and costs","\u002Fdocs\u002Fworkspace\u002Fcosts","docs\u002F8.workspace\u002F2.costs",{"title":799,"path":800,"stem":801},"Partner","\u002Fdocs\u002Fworkspace\u002Fpartner","docs\u002F8.workspace\u002F3.partner",[803,805],{"title":758,"path":759,"stem":760,"description":804,"children":-1},"Publish a workflow as a version and set it up for yourself.",{"title":440,"path":180,"stem":763,"description":806,"icon":764,"children":-1},"Check what an agent wants to write before it happens, and answer its questions during a run.",1791184598077]