Site owners tend to make one of two mistakes. Some jump to a pricier plan at every slowdown and waste money, because the real cause was a heavy plugin or a bot attack. Others refuse to admit the site has grown and suffer under limits for months, losing customers. Here are the concrete signs that an upgrade is due, how to check them and when optimising first is cheaper.
What is limited on shared hosting
Many sites share one server, so each account has limits: disk space, number of files (inodes), concurrent processes, CPU and memory share, database connections and mailbox size. The exact list depends on the plan, so check the plan description or ask support. Upgrading without knowing which limit you hit is a blind decision.
Sign 1: disk or file count near the limit
ISPmanager always shows disk usage. If it stays above 85-90 percent, uploads, updates and backups start failing first. Sometimes the disk has space but no new file can be created, which means the inode limit is exhausted, often by cache plugins and session files. With SSH access, these commands show what takes the space.
du -sh ~/www/* 2>/dev/null | sort -h | tail
du -sh ~/mail 2>/dev/null
find ~ -type f | wc -l
find ~/www -type f -size +100M -exec ls -lh {} \;
Typical culprits are old backup archives stored inside the hosting, forgotten test copies, oversized logs and unread mailboxes. If the disk still fills quickly after cleanup, the growth is real.
Sign 2: 503 and 508 errors at peak times
A site that works all day but fails during a sale or campaign is hitting process or CPU limits, often shown as "Resource Limit Reached" or a 503. Once a month during a big campaign, caching may be enough. Every evening means the site has outgrown the plan.
Sign 3: database errors
"Too many connections" or "Error establishing a database connection" often point to a connection limit, but first look for slow queries or plugins firing dozens of queries per page. If fixing them solves it, no upgrade is needed.
Sign 4: traffic and business growth
The healthiest reason: the site is succeeding. If monthly visits have multiplied and orders grew, the small starter plan naturally becomes tight, and waiting is risky because slowdowns hurt most on your busiest days.
Example: a Tashkent clinic with online booking
A private dental clinic in Tashkent runs misol.uz on WordPress for appointment booking. After a social media campaign the booking form became slow in the evenings, sometimes showing a 508. A review found 6 GB of old backups and 14 unused plugins; they were removed, page cache was enabled and images compressed. The disk problem disappeared, but the evening peak still hit the process limit because the booking form is dynamic and cannot be cached. Only then did the clinic move one plan up, based on real data.
Check before you upgrade
| Sign | Likely cause | Try first |
|---|---|---|
| Disk over 90% | Old backups, logs, mail | Download and delete archives, rotate logs |
| Cannot create files | Inode limit, cache files | Clear cache and session folders |
| 503/508 at peak | Process or CPU limit | Page cache, limit bots |
| Database errors | Heavy queries, connection limit | Find slow queries, fewer plugins |
| Constant slowness | Real growth | Upgrade or move to a VPS |
Always check bot traffic too: blocking an aggressive crawler is cheaper than a bigger plan.
Upgrade or move to a VPS
Upgrading is usually simplest: files, databases and mail stay put and only the limits grow. On sayt.uz you do it from the hosting page in your cabinet, and support can help you choose using your current usage figures. If even the largest shared plan is not enough, or you need root access and guaranteed resources, the next step is a VPS. Either way, base the decision on panel figures, logs and when errors recur.