The invoice for one client says 212 devices. The client’s own count says 198. It is the last Thursday of the month and the person who does the reconciling has a spreadsheet open with two exports in it, one from the monitoring tool and one from the service desk. Fourteen machines were retired in the RMM in the spring and never left the desk. She will find them in about an hour, credit the difference, and do the same thing again next month for the next client. Nobody planned this job. It exists because the two products keep two lists.
Because the two products have two databases. Open a ticket in most service desks and the device attached to it is a copy. Somewhere, overnight, a job read the monitoring tool’s inventory, matched machines by hostname or serial, and wrote the result into the desk’s own table. A sync is the only way to make one product show what the other knows when they were built by different companies, and every vendor who assembled a platform from acquisitions has one of these jobs.
Because the thing it matches on is not stable. Hostnames change. Serials are missing on virtual machines. A device retired in one system lingers in the other because nothing told it to go. A hostname gets reused for a new box and the copy now points at the wrong hardware. Each of these is rare on its own. Across a few hundred devices and a few hundred nights they are routine, and by the time a technician reads the ticket the copy is hours old at best and wrong at worst.
Billing arguments. The invoice says 212, the client says 198, and the difference is fourteen machines that were retired in the RMM but never left the desk. The client is right, and now they know to check.
Wrong tickets. A technician connects to the wrong box because the copy pointed at a hostname that was reused.
Reconciliation time. An afternoon a month per client, done by the person who can least afford it.
The 3 am question. When something breaks, which system was right? If the answer is “neither, exactly”, the post-mortem starts with an argument about the inventory instead of the incident.
The Stack Seams Score words this one plainly: a nightly sync nobody trusts costs one reconciliation a month. That is the visible cost. The invisible one is the client who audits the invoice because they caught the fourteen once.
Not a better sync. Guardian PSA does not mirror devices at all. The device list on a ticket is fetched live from Guardian RMM, because they are the same platform reading the same table. We deliberately retired the sync that used to exist between them. There is nothing duplicated, so there is nothing to drift, and the question of which system is right stops being a question. Here is what the technician sees on a ticket.
The “last check 40 s ago” line is the whole argument. It is not a timestamp on a copy. It is the monitoring platform’s own record, read at the moment the page loaded. Retire the machine in Guardian RMM and it is gone from the desk in the same moment, because there is no second list for it to survive in.
Quite a lot, and none of it is a separate feature. Because the ticket knows the real machine, the intake rule that created it can match on the name of the check that failed, the host group the machine belongs to and the exit code the check returned, not just the words in an email. Because the machine is the real one, the credential for it can be pulled from Guardian Vault and injected into the session server-side, so the technician connects without being shown the password. And because the client portal reads the same table, your customers see the devices you look after for them live, alongside their tickets, contracts and invoices, with what each contact may see switched on one control at a time. What that changes on an ordinary Tuesday.
The device count is right, and so is the time against the ticket, which becomes an invoice at the contract rate and pushes to QuickBooks Online to send and collect. What we do not do yet is generate recurring agreement invoices automatically. Time-and-materials and project billing work today; automated monthly agreement billing is on the roadmap, not in the product. If that is the centre of your month-end, weigh it heavily and raise it on the first call.
Is the device list in your service desk live, or synced overnight? It sounds small. It is the fourth of the ten questions we think actually separate the products in this category, and it is the one that gets the longest silence. A good answer is that the desk reads the monitoring platform’s device records directly, with nothing duplicated. Anything involving the word “nightly” is the reason for the spreadsheet.
The questions that cannot be answered by a feature checkbox, with our own answers, including where ours is a no.