I'd imagine that this is something that's been different from islamic government to islamic government and perhaps hasn't even been exactly the same over time even within the same states

Exactly right. Both those who claim that the jizya was a heavy burden, and those who claim it was no greater than what the Muslims had to pay, are making the same error of imagining that there was some single method of imposing the jizya across all of time and space. In the Ottoman empire, the jizya varied not only in terms of how heavy of a burden it was, but also in terms of how it was collected. In theory, the jizya was meant to be a poll-tax on all non-Muslim males who were capable of paying it, with the amount to be paid adjusted according to whether the individual was "poor," "middling," or "rich." Yet until the reforms of the 1690s, in many places it was actually calculated by household, and not by individual. The level of payment varied a great deal across the empire, and even within the same general region. Ronald Jennings has studied this closely for the first quarter of the seventeenth century, in his studies of the city of Kayseri:

The per capita cizye [jizya] paid in Kayseri increased sharply between 1600 and 1626. In 1605 Kayseri zimmis [non-Muslims] were paying 174 akçe. Between 1612 and 1615 the amount increased to 200 akce; between 1610 and 1623 it increased to 215 akçe, and the following year the level reached 220 akçe. The 26% increase in two decades far surpassed the slight decline in the value of the akçe over that same period. Zimmi cultivators attached to lands belonging to evkaf [pious endowments] paid a slightly lower cizye than did other zimmis. The zimmis connected with the Hacı Ahmed Başa and the Şah Ruh evkaf paid a cizye of 135 akce in 1607, increased by 1615 to 174 akçe, to 189 by 1623, and to 194 by 1626. [...] The cizye in all provinces was not necessarily the same, although the firmans give no explanations of those differences. As early as 1609 and 1610, zimmis of Maraş, Elbistan, and Sis were paying 200 akçe cizye, when 174 akce was still being paid in Kayseri. Although one might have expected to find a graduated tax such as proposed by the great Hanefi jurists Abu Hanifa and Abu Yusuf, the cizye in Kayseri province was discussed strictly on a per capita basis, with no gradients. Although one does not encounter legal cases involving zimmis claiming to be too poor to pay cizye, the apparent method of levying that tax in Kayseri was regressive and may have worked hardship on the zimmi poor. Of course, it is not impossible that within their own communities the zimmis re-allocated the tax burden in accordance with wealth. On occasion the cizye in Kayseri was collected in advance from leaders of the zimmi communities who in turn collected from their communities. [In contrast,] In Karaman in 1583 the cizye was carefully graduated: the rich paid from 60 or 80 to 90 akçe apiece, the poor 30 akçe, and those intermediate paid between 30 and 60 akçe.
[...] The cizye paid by Kayseri zimmis was not innocuous, but it was by no means a crushing burden. For each adult male it represented perhaps half the cost of a donkey. If, as the kanunname of 1500 states, the cizye was the only tax the zimmis paid in excess of what Muslims paid, the amount was certainly not great enough to impose a burden so unequal as to put individual zimmis at a substantial economic disadvantage in their competition with Muslims for land and trade.

Ronald C. Jennings, "Zimmis (Non-Muslims) in Early 17th Century Ottoman Judicial Records: The Sharia Court of Anatolian Kayseri," Journal of the Economic and Social History of the Orient 21 (1978), 232-5.

So, as you can see from this excerpt, the jizya's value could fluctuate in the same place from decade to decade, could vary from region to region, and even the method by which it was calculated and in what way the non-Muslims were expected to pay it was also subject to change. And all this within one part of the Ottoman empire in one very short time-frame. There was no single, uniform jizya applied across Islamic history. To understand the degree to which it was a burden, we would have to look at each situation on a case-by-case basis.