↑↓ selectEnter openEsc close

    1. Elbuz

    Why a price list run failed with an error and how to find the cause

    7 min readUpdated

    When a price list import fails, the step-by-step log shows the failed step and the supplier's site or mail server reply.

    When you need this:

    • The file didn't download — a red log line shows the reply: a 404, a Cloudflare check, a login page instead of the file.
    • No price list for a while — the program knows the supplier's usual interval and flags an error only on a long gap.
    • No time to read the steps — the “✦ AI” button in the source header names the cause and a fix.

    Where to start: the source's “Log” tab → the first red line of the last run.

    What helps you find the cause

    Which step the run stumbled on

    The second-to-last panel of the tab, below the history table, is called “Last run log”, with an “Operations: N” counter next to the title. It is a step-by-step breakdown: what ran and how long it took. This is where you go when a history row says “Error” and the cause is unknown.

    The panel header shows the status badge, the start date and time, the end time, the duration and the file name. If stop rules triggered in this run, a “Rules triggered:” line appears there too — and it shows all of them: both the one that blocked the import and the ones that only warned.

    Last run log panel of a blocked price list: status, date, file name, three triggered rules and the operations table ending with the cancellation row
    1Run status
    2Triggered rules
    3Operations count
    4Import canceled row
    1. The status pill shows the run was blocked by a stop-rule.
    2. The line lists all three stop-rules that fired in this run.
    3. The counter shows how many operations the run logged.
    4. The last row records that the import was canceled and old data kept.

    Next comes the table of steps with the “Time”, “Operation” and “Duration” columns. Error steps are highlighted, and you need the first such row: it names the operation where everything stopped and when. What each import operation does is covered in a separate article on the order of operations; here the point is to see how far the run got.

    Four things worth knowing in advance:

    • Only the first 300 steps are shown. If there are exactly that many, the bottom line reads “Showing the first 300 operations of the run.” — for the rest, go to the Operations Center.
    • All steps are shown in a row, including service ones. There is no filter by severity here, so the list is long, and that's normal.
    • The panel is limited to this price list. When you import a dozen sources at once, steps from the other sources don't end up here.
    • Sometimes one more column is added — “Cost”. It appears only when the run included a paid AI step. A regular price list import never has it.

    Sometimes the panel is empty and says “No log records for the last run”. The run figures in the table above are still correct — there are simply no step-by-step records saved for this particular import. The breakdown will appear after the next import.

    The file didn't download: what the supplier replied

    When the file couldn't be fetched by link or from email, the program doesn't write a vague “it failed” to the log, but exactly what the site or mail server replied.

    This line comes right after “Getting file by link” (or after the step that fetches the file from email), highlighted in red, and the whole run gets the “Error” status. The new file isn't used: the run continues on the previous file, and the products in the price list stay as they were.

    Last run log panel: an Error run whose first step is getting the file by link, followed by a red line with the supplier's 404 response
    1Error status pill
    2Red line with the 404 cause
    3Reading data from file
    1. The Error pill shows that the last run ended with an error, with its date, duration and file name.
    2. The red line right after getting the file by link says the supplier site answered with code 404 and the file was not downloaded.
    3. Reading data from file continues on the previous file, so the products from the earlier download stay in use.

    For a link source, the program checks what came back in this order and stops at the first cause:

    1. Connection failed — “Could not download the file by the link: ‹error text›. Products from the previous file are in use.” The supplier's server didn't respond, the address doesn't exist, or the connection dropped.
    2. A Cloudflare browser check instead of the file — “The supplier site showed a browser check (Cloudflare) instead of the file: the program downloads price lists without a browser and cannot pass it.” The same line then says what to do: ask the supplier for a direct link without protection or to whitelist our server, or receive the price list by email or .
    3. Response code 400 or higher — “The supplier site responded with code ‹code› — ‹explanation›. The file was not downloaded, products from the previous file are in use.” The explanation depends on the code: 401 and 403 — “access denied: login is required, the key in the link expired, or the site blocks our server”; 404 and 410 — “file not found by the link: the supplier may have changed the address”; 429 — “the site limited the request rate, the import will retry on schedule”; 500 and above — “the supplier server failed or is unavailable, usually temporarily”; anything else — “the request was rejected”.
    4. An empty file — “The link returned an empty file. Products from the previous file are in use.”
    5. A web page instead of the file — “The link returned a web page instead of a price list file (for example, a login page or an error page).” This happens when the site serves a login or error page with a success code. An HTML page with a table is accepted as a file: that's what the “xls” exports of some accounting programs look like, and Excel reads them.

    For email, there are two such lines:

    • “Could not connect to mailbox ‹address› (server ‹server›): ‹reason›. The new email was not checked — check the login, password and mail server settings.” — the mailbox didn't let the program in;
    • “Mailbox ‹address› (server ‹server›) returned an error while searching for the email: ‹reason›. This is usually a temporary mail server failure — the import will retry on schedule.” — the connection worked, but the server failed during the search.

    In both cases, the file left over from the previous email isn't used: before 18.10.2026 it was picked up silently, and old prices looked like a fresh import.

    XML price lists can get a separate line if the file arrived but contains a site page instead of : “The price list file is a web page (HTML), not XML: a site page came instead of the price list — a browser check (Cloudflare), a login page or an error page. Check the link or the email attachment. Parse error: ‹text›”. It applies to any delivery method, including an email attachment.

    No new file: warning or error

    A supplier who sends a price list once a week has no new file six days out of seven — and that's normal. So a run with no new file, where the source still responded without failures (the mailbox simply has no new email), isn't marked as an error.

    The log shows “No new file received from the supplier — products from the previous file are in use. This is not an error if the supplier has not sent an update yet.”, then “Reading data from file: ‹name›”. The line isn't red; if the run has no other red lines, it ends as “Done”, and products are updated from the previous file.

    If the source's history already has a run with a new file, the next line notes when a new price list last arrived and how often it usually does, for example: “Last new price list: 2026-10-11 (7 days ago). The supplier usually sends a new one every 7 days.” The second sentence comes in three variants:

    • “The supplier usually sends a new one every N days.” — the program worked out the interval from the source's history;
    • “By the price list setting, a new file is expected every N days.” — a specific frequency is selected in the source card;
    • “There is little import history yet — the program expects a new file at least every 7 days (you can change this in the price list card, Schedule section).” — the source has fewer than three new files so far.

    A run becomes an error only when there has been no new file noticeably longer than usual. Then instead of the note there's a red line: “No new price list for N days (last one: ‹date›). ‹how often it's expected› Check that the supplier is still sending it and that the address, email subject or file link have not changed.”

    The threshold is the larger of two: the interval plus one day, or one and a half intervals. For a daily price list that's over 2 days, for a weekly one over 11, for a biweekly one over 21, for a monthly one over 45.

    A few more details so the log doesn't surprise you:

    • The note and the overdue line are also written when a file formally arrived but is byte-for-byte identical to the previous one and the price list already has products — for example, the link returned the same price list. In that case there's no “No new file… received” line.
    • A price list uploaded manually as a file isn't checked: nobody sends such a file.
    • If fetching the file already ended with a red line — a link or email failure (see above), an FTP error, a closed Google file, a date macro in the link, missing email access — there's neither a warning nor a note: the cause is already named, and the run is marked as an error.
    • If “Do not monitor” is selected in the source card, only the warning remains, with no note and no overdue line.
    • Runs before 18.10.2026 show “No new file received, using the previous one” instead of the warning, and such runs are marked as errors — that's how the program used to behave.

    How to choose the frequency for a specific supplier and how the program learns from the history — see the article on connecting a source: how often the supplier sends a new price list.

    How to get a breakdown with the “✦ AI” button

    In the header of the selected source, next to “Update my catalog”, there's a “✦ AI” button. It opens the consultant chat, which reads the same records as the “Log” tab: the last run, its report and the log steps. In the “All sources” mode the button isn't there — select a source first.

    If something is wrong with the last import, the button shows an orange number, and the hover tooltip names the problem:

    • “a stop rule stopped the last load — new data was not applied”;
    • “the last load failed — the previous product list is in use”;
    • “stop rule warnings fired in the last load”;
    • “formulas with errors were skipped during the load”.
    Header of the blocked source Price Ceres.xlsx with the AI button and a badge showing 1 item
    1AI button with badge
    1. AI button in the source header — a badge shows the number of items the assistant can analyze.

    Clicking the button with a badge sends the breakdown right away: the consultant names the cause from the run report and one step to fix it. After that the number turns gray; clicking again only opens the hints. If the set of problems changes, the number turns orange again.

    AI consultant analysis of why the Ceres price list import was blocked
    1Cause and fix
    1. The AI consultant names the cause (stop-rules triggered by 2,223 skipped rows) and the fix.

    Without a badge, the button opens hints about this source: “What changed in the last load” (new and missing products, no price, price changes, stop rules), “Set up columns / XML branch” (for Excel, , XML and ), “How this price list takes part in price choice” (if the price list products are matched with the catalog) and “Price list field formula”.

    AI consultant window after a new chat and a repeated click on «AI»: record card with hint buttons
    1First hint: the stop-rule that blocked the load
    1. Hint buttons under «About this record» let you ask typical questions about the source in one click.

    The consultant changes nothing on its own: it proposes a mapping or schedule change as a plan with a confirm button. Phrases for the consultant are in the article about the AI consultant.

    Scenarios

    The supplier's site responds with 403 or 404

    What to do. Check the link or the key in it with the supplier and replace the link in the source card.

    Result. The next run picks up the new file. Until the link is fixed, every run is marked as an error, and the price list uses the products from the previous file.

    The site is behind a Cloudflare check

    What to do. Ask the supplier for a direct link without this protection or to whitelist our server. Another option is to receive the price list by email or from FTP.

    Result. The program can't get past the browser check: it downloads the file without a browser. After any of the three steps, the price list imports again.

    The supplier sends a price list once a week

    What to do. Nothing, as long as the line “No new file received from the supplier” is followed by a note rather than an overdue line. If the supplier sends less often than the program thinks, select the frequency in the source card, “Schedule” section.

    Result. Runs without a new file end with the “Done” status, and only a genuinely overdue file becomes an error.

    What to do about a run error

    Run errors in the step-by-step log
    What you seeWhyWhat to do
    The “Error” badge in a history row At least one step of this run ended with an error Expand “Last run log” and find the highlighted row — it has the exact cause (step-by-step log)
    “No log records for the last run” No step-by-step records were saved for this particular import The run figures in the table above are still correct; the breakdown will appear after the next import
    “Showing the first 300 operations of the run.” The run has more steps than the panel limit The first 300 steps include both reading the file and the main operations; for the rest, go to the Operations Center (step-by-step log)
    “No data received, previous product list will be displayed” The source returned nothing — most often it's the or the link Check access to the source; this run's figures belong to the old data, not the new file
    “The link returned a web page instead of a price list file…” or “The price list file is a web page (HTML), not XML…” Instead of the file, the site served a login, error or browser check page Open the link in a private browser window: it should download the file right away (causes on the supplier's side)
    “The supplier site responded with code 403 (404, 429, 5xx) — …” 403 — no access, 404 — no file at that address, 429 — too many requests, 5xx — the supplier's server failed 403 and 404 — check the link or the key in it with the supplier; 429 and 5xx usually clear up on their own, the next run follows the schedule (response codes)
    “The link returned an empty file. Products from the previous file are in use.” The server responded but returned nothing Check the link in a browser; if the file is empty there too, it's a question for the supplier
    “Could not download the file by the link: ‹error text›…” Couldn't connect to the supplier's server Check the address; if it's correct and the site works, retry the import later
    “Could not connect to mailbox…” or “Mailbox … returned an error while searching for the email…” The mailbox didn't let the program in (login, password, server), or the mail server failed during the search For the first, check access to the mailbox; the second is usually temporary, and the import will retry on schedule (about the mailbox)
    “No products found in node “…” — check the XML product branch setting” The product branch points to the wrong node, or the supplier changed the feed structure Fix the product branch in the “XML structure” section of the source card
    “No file found to import” The file is neither in the source nor in storage Upload the file or set up how it's fetched (about uploading a price list)
    “Column addresses for loading data from the price list are not filled in!” The columns in the source card were left unmapped Map the columns in the “Columns” section of the card (about uploading a price list)
    “Column addresses are not filled in for columns flagged “Consider uniqueness”!” The field is marked as unique, but its address in the file isn't set Set the column address or clear the uniqueness checkbox
    “Import interrupted by the user!” “Stop” was clicked Nothing: the run goes to “Canceled” (history table)
    “Background loading was force-stopped (processes terminated: …)” “Stop” had to be applied forcibly — the process didn't stop by itself Nothing; the next run works as usual
    “Aborted: the process crashed or showed no activity” The process broke off, and the safeguard closed the operation Run the import again; if it keeps happening, look at the file size and timeouts
    “This source is already being imported (started at …). Wait for it to finish or stop it with the “Stop” button.” Two imports of the same source can't run at the same time Wait for it to finish or stop it (about starting an import)
    “Failed connect to ‹host›:21; Connection timed out” The supplier's FTP server is unavailable; the text comes from the server itself Check the credentials and the address — it's not a program error

    FAQ

    The status is “Error”, but the price list has products. How?

    Either the error came after part of the work was done, or no new file arrived and the program kept the previous product list. In the second case, a red line with the cause follows the file-fetching step (causes). The products are old, even though the import date is today.

    The price list is collected by a parser from the supplier's site and imports nothing. Where are the details?

    The run summary is here, on the “Log” tab. What happened to each page of the site is shown in the “Parsing log” window: source header → “Site data import settings” → “Additional” → “Parsing log” (how to read it).

    The log steps are in Russian, although the interface is Ukrainian. Can they be translated?

    No. A step's text is saved in the session language of whoever started the run, and steps already written aren't translated. Runs started from the Ukrainian interface will be in Ukrainian.

    Do steps from other sources end up in this log?

    No. The panel is limited to this price list: when a dozen sources import at once, steps from the others don't end up here.

    Where does the “Cost” column in the log come from?

    It appears only when the run included a paid AI step. A regular price list import doesn't have it.

    Where to find it and who has access

    Menu
    Directories → Processing price lists → List of price lists → source → “Log” tab
    Other ways in
    the “Log” button on a failed import card in the Operations Center; the “✦ AI” button in the source header
    Access right
    the “Price lists” window (in the menu — “List of price lists”)

    Related topics

    Was this article helpful?

    Missing a feature? Tell us.

    We will add it for free if other users need it too.

    If only you need it, we will agree on it separately.

    Article contents

    Suggest an improvement

    Try Elbuz free