My battery pack is old, but still ticking away. They are sealed AGM types, 24 cells at 2 volts each, designed for commercial building backup power supplies. I got them at the "end" of their life, but the standard of reliability for a national ISP service is much higher than my own, so I've benefited from 5 years of service from them. When I got them they were in a very sorry state. Broken bus bars, dropped in an alleyway, on a rainy day. After cleaning them up and keeping them in an insulated (but not heated) building for years, they have delivered the juice consistently ever since. No leaks, no off-gassing, no bulges on the front (though they are swelled enough inside their steel containers that I can't slide the individual cells out).
That said, they DO show signs of deterioration. It's probably worth posting about because I've come up with a method of monitoring my batteries that I haven't seen written about. For any forum users that have a dozen or more cells, this might help to spot a bad trend. Before it becomes a disaster.
It's likely that many owners of batteries monitor them with voltage measurements, and I expect they measure cell by cell, looking for cells with voltage lower than the average. I've been doing that fairly regularly, too (if every 4 months counts as "regularly"). Since I write the results down as I go, I've started entering the numbers into an Excel spreadsheet. Now the trends are obvious. Cells 6 and 8 are consistently the worst in the pack, while cells 10 and 16 are usually low, but not always. Other cells tend to be very good, such as 7, 13, and 18.
Then I tried a "standard deviation" on the 24 cells in my stack. That's when another overall trend showed up. I was able to go back through the measurements I recorded years ago, and get a standard deviation for the cells on each date. Plotted on a chart, the trend is clear: the variation between the cells is growing, meaning the bad are getting worse.
On the graph, the variability between the cells was very high when I got them; off the chart. Considering the mess they were it, that's expected. After repeated re-charging from a 2-cycle (bulk/float) battery charger the standard deviation went way down, and I was able to keep it down after installing them and keeping them with wind power. The variability still went up and down from time to time, and then in 2010 I put solar panels on the system. From then on the variability has followed a continuous trend, creeping upward ever so slowly and inexorably. The bad cells are getting badder.
To make sense of the numbers, it means that my 4 worst cells are usually 0.0065 volts below the average or lower. Actually, my worst cell (#8) is 0.018V below the average today. On cells whose standard voltage reference (2V) should be 2.25 volts at float, this cell is undercharging by 7%. It's like a fully-charged 12V battery resting at 12.55V rather than 12.6V. Showing its age.
On the whole, the group seems to be in OK health, but the bad cells are going to kill the batch someday.